Downstream Environmental and Socio-Economic Impacts of the Ranganadi Hydel Project in Northeast India

A THESIS SUBMITTED TO

MANIPAL ACADEMY OF HIGHER EDUCATION

FOR FULFILLMENT OF THE REQUIREMENT FOR THE AWARD OF THE DEGREE

OF DOCTOR OF PHILOSOPHY

BY PRIYAM LAXMI BORGOHAIN

UNDER THE GUIDANCE OF DILIP R. AHUJA

National Institute of Advanced Studies Bengaluru, India

2019

Dedicated to my parents

Acknowledgements

I would like to express my deepest respect and gratitude towards Dr. Dilip R. Ahuja, my PhD supervisor, for his guidance, innumerable suggestions, constructive criticism, and moral support since the conception of this work.

I am also hugely grateful to my SAC members, Prof. M. Amarjeet Singh and Dr. M. Mayilvaganan, for their invaluable comments and advice, that helped me give this thesis its present form.

I thank the Department of Science and Technology, India, for the financial assistance in the form of research fellowship under the DST INSPIRE Fellowship program.

I would also like to take this opportunity to thank Mr. P. Srinivasa Aithal, Mr. A. Devaraju, Ms. Hamsa Kalyani, Ms. R. Vijaylakshmi, Mr. V.A. Ramesh, Ms. V.B. Mariyammal and Ms. S. Lalitha, for all their help and assistance in NIAS. I am forever indebted to Ms. J.N. Sandhya, for her guidance and support through all the academic and administrative procedures. I also thank Dr. Hippu S.K. Nathan for sharing his expertise and constructive comments.

Fellow scholars and colleagues, who go on to become close friends, are the ones who can truly understand and appreciate the PhD journey. I am extremely thankful to Neesha, Ankita and Sanket, for their love and friendship, the never-ending discussions, agreements and disagreements, motivation and support throughout my PhD journey. I also thank each and every person who has been my hostel mate, and with whom I had shared the badminton court in NIAS, for providing a memorable and fun-filled stay in campus. I especially thank Madeni and Julie, for giving me a home away from home in Bangalore.

I am thankful to all those who helped me conduct my fieldwork. A special thanks to Bonti Borma, who did not hesitate to walk for miles under the sun and rain as she accompanied me on the field and for giving me a loving home during fieldwork. Most importantly, I thank the residents of my study villages in Lakhimpur district, for trusting me with their views and opinions.

I have been immensely fortunate to be gifted with a second family - my in-laws, Mr. Ratna Kanta Hazarika, Mrs. Bijoya Hazarika and Babi Hazarika. I thank them for the love,

iv

Acknowledgements understanding and unwavering support that enabled me to finish this thesis, and for giving as much value to my work as I did.

This thesis would not have seen the light of day had it not been for the constant support and encouragement from my better half, my best friend and my husband - Dr. Pranjit Hazarika. I take this opportunity to thank him for his immense patience, guidance and all the sacrifices he made to help me finish my work. He has been the biggest supporter and the greatest critic of this thesis, and I am grateful to him for all the long discussions and arguments, and for lending a patient ear to my deductions, despite the subject being outside his expertise. Above all, I thank him for never letting me to give up on myself.

I am also grateful to my brother - Dr. Manas Krishna Borgohain and sister-in-law - Dr. Deepsikha Saikia, ‘My PRECIOUS!!!!’, who put up with all the drama, yet never stopped believing in me. I thank them for having more faith in me than I had in myself, and motivating me at every step of this work.

Last, but not at all in the least, I thank the two greatest cheerers and teachers in my life, my parents - Dr. Krishna Kanta Borgohain and Mrs. Gopa Borgohain. Words fail to express my gratitude towards them, who never stopped believing in me and picked me up when I hit the lowest point in this journey. They egged me on, even when I wanted to give up, and none of this would have been possible had it not been for their love, support and encouragement. Therefore, I dedicate this thesis to them.

v

Abstract

The Brahmaputra Basin in Northeast India is identified as the ‘future powerhouse’ of the country, given its vast hydropower potential of 66 GW. Consequently, the region is now witnessing rapid development of hydropower projects. At the same time, large dam construction faces strong opposition given their adverse environmental and socio- economic impacts. Arunachal Pradesh has the highest number of upcoming projects in the basin. One of the major concerns of dam building in this Himalayan state is their downstream impacts upon the riverine and riparian ecosystems of the lower floodplains of Assam.

Commissioned in 2002, the 405 MW Ranganadi hydel project (RHEP) is the sole operating hydropower scheme in Arunachal Pradesh. As an inter-basin water diversion project, it involves transfer of flow from the impounded Ranganadi River to an adjacent river – Dikrong, for power generation. This study analyzes the impacts of the hydel project upon the downstream hydrology and geomorphology of Ranganadi and Dikrong within the floodplains of Assam. The study further examines the associated socio- economic impacts of impoundment upon the floodplain communities of selected downstream villages in the Ranganadi basin.

The two affected rivers displayed contrasting patterns of hydrological and morphological changes in the post-dam period. Annual median flows in the impounded and flow-deprived Ranganadi decreased by 63%, while the reduction in monthly medians ranged from 46% to 80%. The annual extremes showed higher reduction in the minimum water conditions. Unlike the pattern of change commonly demonstrated by impounded rivers, Ranganadi displayed attenuation of both high and low flows, indicating the influence of unsustainable water diversion for off-site electricity generation. At the same time, the flow-recipient river - Dikrong displayed 138% increase in post-dam annual flows. The monthly flows increased significantly by more than 30%. Dikrong, too, exhibited higher elevation in low flows compared to monsoonal high flows.

Ranganadi exhibited decreased braiding and a temporal narrowing channel pattern, while Dikrong displayed channel widening and an increasingly braided and multi-channeled planform. Sudden water releases from the reservoir, flash floods and sand casting were the major problems perceived by the downstream communities in the Ranganadi basin. Farming as a primary livelihood decreased, particularly in the left bank

vi

Abstract

villages and especially post-2008 flash floods. The major adaptations observed were livelihood diversification and increased leasing of agricultural lands.

The downstream impacts of river damming have been aggravated by a complex overlapping of natural and anthropogenic stressors, and would become more unpredictable in the future given the cascade development of dams in both river basins. The study found the current flow release pattern of RHEP to be unsustainable, where downstream flows especially in Ranganadi have been mismanaged. Finally, this study emphasizes the recognition of downstream impacts in the Environmental Impact Assessment framework and a space for the downstream riparian to negotiate their concerns. It also recommends the implementation of environmental flow releases by old and new projects alike, especially in cases of inter-basin water diversion.

Keywords: Hydropower, downstream impacts, Ranganadi, Dikrong, hydrology, morphology, socio-economic

vii

Contents

Declaration i

Certificate ii

Acknowledgements iv - v

Abstract vi - vii

List of tables and figures xii - xx

Chapter 1: Introduction 1 - 24

1.1 Introduction 1 1.2 Study area 6 1.3 Literature review 8 1.3.1 Energy scenario in India and the rationale behind 8 hydropower development 1.3.2 Northeast hydropower scenario 12 1.3.3 Downstream impacts of dams 14 1.3.4 Impacts on downstream river hydrology and geomorphology 15 1.3.5 Downstream socio-economic impacts 16 1.4 Statement of the problem 20 1.5 Why the Ranganadi hydel project? 21 1.6 Objectives 22 1.7 Chapter outline 22

Chapter 2: Hydrological and Morphological Changes in River 25 - 68 Ranganadi

2.1 Introduction 25 2.2 Study area - Ranganadi basin 26 2.3 Data and methodology 27 2.4 Results and discussion 33 2.4.1 Alterations in river hydrology 33 2.4.1.1 Group 1 alterations 36 2.4.1.2 Group 2 alterations 37 2.4.1.3 Groups 3, 4 and 5 alterations 39 2.4.1.4 Changes in high flows and floods 40 2.4.2 Changes in river morphology 43

viii

Contents

2.4.2.1 The avulsion of 1991 and its repercussions on channel 43 morphology 2.4.2.2 Channel sinuosity and braiding 43 2.4.2.3 Disappearance of side channels across R5, R6, and R7 47 within reach 2 2.4.2.4 Changes in channel width 48 2.4.2.5 Absence of a visible water signature in the downstream 52 reaches of the 2008 channel and its impacts on wetted channel width 2.4.2.6 Channel width calculated up till section R26 53 2.4.3 Bankline migration, erosion, and deposition patterns 54 2.4.3.1 Overall bank migration between 1987-2002 and 56 2002-2014 2.4.3.2 Periodic bankline migration 60 2.4.4 Capture of the Joyhing channel and diversion of flow in the 63 post-dam period 2.4.5 Comparison between flow conditions and channel width 66

Chapter 3: Hydrological and Morphological Changes in River Dikrong 69 - 103

3.1 Introduction 69 3.2 Study area - Dikrong basin 70 3.3 Data and methodology 72 3.4 Results and discussion 75 3.4.1 Alterations in river hydrology 75 3.4.1.1 Group 1 alterations 77 3.4.1.2 Group 2 alterations 78 3.4.1.3 Groups 3, 4 and 5 alterations 80 3.4.1.4 Changes in high flows and floods 80 3.4.2 Changes in river morphology 82 3.4.2.1 Channel sinuosity and braiding 82 3.4.2.2 Changes in channel width 85 3.4.3 Bankline migration, erosion, and deposition patterns 90 3.4.3.1 Overall bank migration between 1987-2002 and 90 2002-2014 3.4.3.2 Periodic bankline migration 96

ix

Contents

3.4.4 Comparison between flow conditions and channel width 100

Chapter 4: Socio-economic impact of the Ranganadi hydel project on 105-146 downstream riparian populations in the Ranganadi floodplain

4.1 Introduction 105 4.2 Data and methodology 106 4.3 Background of the study villages 108 4.4 Key geographical features of the study villages 112 4.5 The impact of June 2008 flash flood 115 4.6 Results and discussion 117 4.6.1 Impact on agricultural lands and the resultant changes in 117 cultivation 4.6.2 Impact on the livelihood structure of the downstream 121 communities 4.6.2.1 Changes in primary livelihood activity 121 4.6.2.2 Livelihood diversification and adoption of multiple 123 livelihoods 4.6.2.3 Impact on downstream fish availability and on the 125 practice of fishing 4.6.2.4 Impact on animal husbandry including grazing 129 patterns 4.6.3 Changes in riverine use and dependence 130 4.6.4 Perceptions of the downstream populations of the hydel 134 project 4.6.4.1 The Subansiri anti-dam movement and its influence 141 upon opinion formation towards RHEP 4.6.5 Variables influencing riverine dependence and the spatial 141 distribution of impacts

Chapter 5: Understanding the complexities around Ranganadi Hydel 147 - 183 Project and broader hydropower development in Arunachal Pradesh

5.1 Introduction 147 5.2 Discussions 148

x

Contents

5.2.1 Comparing the post-dam changes between Ranganadi and 148 Dikrong 5.2.1.1 Hydrological changes 148 5.2.1.2 Low flow alteration 149 5.2.1.3 Monsoonal flow alteration 150 5.2.1.4 Morphological changes 152 5.2.1.5 Sequence and pathway of downstream changes 153 socio-economic impacts 5.2.2 Floods and RHEP’s involvement 154 5.2.3 Complexities surrounding RHEP 159 5.2.3.1 The problem with embankments 159 5.2.3.2 Lack of floodplain zoning 160 5.2.3.3 Cascade development of dams 160 5.2.3.4 Lack of adequate data and baseline information 164 5.2.4 Status of hydropower development in Arunachal Pradesh 166 and issues therein

5.2.4.1 Hydropower development in the Subansiri basin 169 5.2.4.2 ‘Run-of-river’ - a misleading categorization 172 5.2.4.3 The involvement of private players and controversies 175 therein 5.2.5 Environmental Impact Assessment of hydropower 178 development in India - disadvantages of the downstream riparian

Chapter 6: Summary and Conclusions 185 - 193

6.1 Summary of thesis findings 186 6.2 The way forward 192

References 195 - 216

Appendices 217 - 220

Appendix A1 217 Appendix A2 219

xi

List of tables and figures

Tables

1.1 The basin-wise status of hydro-electric potential development in India in 4 terms of the total installed capacity (as on 31.05.2017)

1.2 Features of the Ranganadi Hydel Project 9

1.3 All India Installed Capacity of Power Stations (Utilities, as on 11 31.03.2017) (in MW)

2.1 Annual flow statistics of the Ranganadi derived from daily discharge data 35 from 1991 to 2014

2.2 Difference between the pre-dam and post-dam flow regimes of the 38 Ranganadi following construction of the RHEP, obtained from IHA analysis of daily flow data from 1991 to 2014. The asterisks, ‘*’ and ‘**’, indicate significant differences between the two periods at 0.05 and 0.01 confidence levels (Mann Whitney significance test)

2.3 Temporal and spatial variations in the sinuosity indices (SI) and braid- 47 channel ratios (B) measured across the 5 reaches of the Ranganadi River as well as overall examined length from 1973 and 2014

2.4 Average, maximum and minimum bankfull and wetted widths of the 51 Ranganadi from 1973 to 2014 obtained from 33 sections along the study length of the river

2.5 Overall bankline migration, erosion and aggradation between pre-dam 57 (1987-2002) and post-dam (2002-2014) time-periods

2.6 Periodic bankline migration, erosion and deposition patterns of the 62 Ranganadi between 1987 and 2014 along with the most affected channel sections.

3.1 Annual flow statistics of the Dikrong derived from daily discharge data 77 from 1987 to 2014

xii

List of tables and figures

3.2 Difference between the pre-dam and post-dam flow regimes of the 79 Dikrong following completion of the Ranganadi hydel project, obtained from IHA analysis of daily flow data from 1987 to 2014. The asterisks, ‘*’ and ‘**’, indicate significant differences between the two periods at 0.05 and 0.01 confidence levels (Mann Whitney significance test)

3.3 Temporal and spatial variations in the sinuosity indices (SI) and braid- 83 channel ratios (B) measured across the 5 reaches of the Dikrong as well as overall examined length from 1973 and 2014

3.4 Average, maximum and minimum bankfull and wetted widths of River 86 Dikrong from 1973 to 2014 obtained from 35 sections along the study length of the river

3.5 Overall bankline migration, erosion and deposition of Dikrong between 91 pre-dam (1987-2002) and post-dam (2002-2014) time-periods

3.6 Periodic bankline migration, erosion and deposition patterns of the 97 Dikrong between 1987 and 2014 along with the most affected channel sections

4.1 Codes assigned to the downstream villages that were surveyed along the 108 two banks of the Ranganadi and the number of households interviewed in each village

4.2 Salient demographic features of the six surveyed villages located along 111 the Ranganadi. The distance of the village from the river and the dam has been approximated from GPS points collected during the field survey and Google Earth. The demographic information is as per the Census of India (2011) data (District Census Handbook, Lakhimpur)

4.3 Village-wise distribution of households (HHs) according to their primary 121 livelihood activity between pre-dam and post-dam periods (in percentage). The primary livelihood activity has been broadly distinguished into 3 categories based on whether a particular household is engaged primarily in floodplain agricultural activities, or other non-farm and wage/casual labor livelihoods 4.4 Village-wise distribution of households (in percentage) who have 126

xiii

List of tables and figures

practiced fishing in the riparian ecosystem either prior to or post-dam construction and the changes in riverine and riparian fish availability post- dam according to their observations (Changes in fish availability have been opinionated by HHs irrespective of whether they practice fishing or not).

4.5 River use and dependence of the riparian communities in the sample 131 villages before and after dam construction

5.1 Differences in flow alteration between Ranganadi and Dikrong after 151 commissioning of the RHEP

5.2 Hydroelectric projects planned in the Ranganadi and Dikrong basins 161 besides the existing Ranganadi Hydel Project (RHEP) in Arunachal Pradesh

5.3 Status of hydropower development in the Northeastern states of India 167 (as of 2017)

Figures

1.1 Study area showing the two concerned rivers - Ranganadi (impounded 7 river) and Dikrong (flow-recipient river), (a) location of the dam and reservoir in the former, (b) location of the project’s powerhouse in the latter, (c & d) starting points of the study lengths in each river

2.1 Slope variation in Ranganadi 27

2.2 (A) The Ranganadi basin showing the river and the study length and 27 location of the gauge station, (B) the diversion dam and reservoir as seen in Google Map and, (C) photo of the dam and reservoir collected during field visit

2.3 (A) Channel features that were distinguished following digitization of the 30 river and how the bankfull and wetted channel widths of the river have been defined, (B) the 33 transects and 5 reaches that were distinguished to determine planform characteristics such as width, sinuosity and braiding

xiv

List of tables and figures

2.4 Right banklines of the Ranganadi channel in 2002 and 2014, where the 32 grey polygons indicate the area of migration between the two examined years

2.5 Pattern of daily flow in Ranganadi from 1991 to 2014 observed at the 34 Pahumaraghat G/D station, Lakhimpur District, Assam

2.6 (A) Pre-dam (1987 to 2002) and post-dam (2003 to 2014) hydrographs of 34 daily flow in Ranganadi (Pahumaraghat gauge station) displaying the difference between the two periods, (B) annual flow duration curves for pre-dam and post-periods obtained from daily flow

2.7 IHA parameters displaying the difference between the pre-dam and post- 37 dam periods with respect to the - (A) median monthly (i.e., Group 1 parameters) and (B) annual extreme water conditions (i.e., Group 2 parameters)

2.8 Temporal variation in the median high pulse duration (A), rise rate (B) 40 and fall rate (C), of water conditions in Ranganadi between the pre-dam and post-dam periods

2.9 Pattern of decrease in median annual flow in Ranganadi from 1991 to 41 2014

2.10 Variation in annual peak flows from 1991 to 2014, displaying the distinct 42 lowering from 1999 to 2007 (time-scale 2), i.e., a few years prior to and immediately following impoundment by RHEP

2.11 (A) Breach in the right bank of the Ranganadi channel below section R26 44 in 1991, (B) subsequent avulsion and formation of a new channel in 1995, (C) the large migration polygon formed due to the shift in the banklines of 1991 and 1995 as a consequence of this avulsion, (D) 2002 image showing the abandonment of the old channel and diversion of flow into the new channel on the west, (E) activation of the original channel in 2006 and (F) reformation of the large migration polygon between the banklines of the 2002 and 2006 channel as a result of this return

2.12 The side channels between sections R5 and R7 that show a distinct water 45

xv

List of tables and figures

signature in the post-monsoon images of 2002 and 2010 (A and B), but are not clearly visible in the images of 2013 (C) and 2014 (D), suggesting lack of adequate flow in the river

2.13 Lower reaches of the Ranganadi - reach 4 and 5, displaying a sinuous 46 channel pattern

2.14 Temporal variation in the bankfull (A) and wetted (B) channel widths of 50 the Ranganadi from 1973 to 2014

2.15 Temporal variation in the average bankfull width across the five reaches 51 in Ranganadi from 1973 to 2014

2.16 The fragmented water channel from section R17 to downstream that 53 shows the river to be dry and devoid of flow in 2008 (post-monsoon image dated 26th November)

2.17 The large migration polygon that was generated during the examined 55 periods of 1987-2002 (pre-dam) and 2002-2014 (post-dam) due to the 1991 avulsion and subsequent shift of primary channel below section R26. Both banks of the river, thereby, shifted to the west and remained so till their return back to the original path of flow in 2006. The avulsion polygon between 1987 and 2002 indicates the westward shift of banklines, while the one between 2002 and 2014 indicates an eastward shift of banklines due to the aforementioned return

2.18 Bankline migration patterns in Ranganadi during the pre-dam and post- 59 dam periods, (A) Channel straightening across R11 that resulted in the largest migration by the left bank (in a westward direction) between 1987 and 2002, (B) mirroring of meander bends across R19 and R20 resulting in aggradation along the left floodplain due to westward shift in the former and aggradation in the latter due to eastward shift of the left bankline, (C) prominent shifting of the left bankline to the west from R5 to R7 during the post-dam period (i.e., 2002-2014) due to partial abandonment of the side channels

2.19 Temporal and spatial transformation of the Joyhing channel that joins the 64 Ranganadi on its left bank at two points (C1 and C2). The satellite images

xvi

List of tables and figures

shown in the figure belong to 1987, 2000, 2010 and 2014 (scale = 1:40000) and it can be observed that the Joyhing is much smaller in size and not clearly visible in the pre-dam images (1987 and 2000), whereas the same can be clearly distinguished in the post-dam images of 2010 and 2014.

2.20 The dry channel bed of the Joyhing (A) and the bamboo bridge (B) that 65 connects the mid-channel island bar, caught between the channel and Ranganadi, with the left bank floodplain of the latter and the district headquarters. The images shown in the figure were taken during field visit in December, 2013 and the bridge is located near No.1 Pachnoi Ujani village on the island.

3.1 Slope variation in the Dikrong basin. 70

3.2 Reaches and transects along the river across which channel width, 74 sinuosity and braiding were measured.

3.3 Annual flow duration curve obtained from daily flow in Dikrong at 76 Sissapathar gauge station from 1987 to 2014.

3.4 Flow regime characteristics of the Dikrong before (1987-2002) and after 76 (2003-2014) dam operation, (A) pre-dam and post-dam hydrographs of daily flow, (B) annual median flow.

3.5 IHA parameters showing the difference between the pre-dam and post- 80 dam flow conditions - (A) Group 1 (median monthly flow conditions), (B) Group 2 (median annual extreme water conditions).

3.6 Group 4 and 5 IHA parameters that showed significant difference in pre- 81 and post-dam water conditions, (A) duration of low pulses (in days), (B) fall rate, (C) no. of reversals.

3.7 Transformations in channel planform in Dikrong between 1973, 1987 and 84 2014, (A) increased development of braid bars and multiple channels along reaches 2, 3 and 4 in 1987 contrary to a predominant single-channel meandering planform in 1973, (B) continued increase in braiding and elimination of meanders in reaches 4 and 5 from 1987 to 2014.

xvii

List of tables and figures

3.8 Temporal variation in the average bankfull channel width along the five 86 reaches from 1973 to 2014.

3.9 Temporal variation in the overall channel widths in Dikrong measured 87 over 33 transects along the river’s study length, (A) bankfull channel width, (B) wetted channel width.

3.10 Planform changes that caused width increases during different years, (A) 89 increase in sectional widths between D21 and D29 in 1991 due to avulsions and secondary channel formation, noticeably at D27, (B) changes at sections D23 and D25 in 2008 that were a repetition of past changes in 1991, (C) initiation of a secondary channel on the right bank between D15 and D19 that grew into two prominent channels running near parallel but bending away from each other by 2014, thereby increasing the sectional width at D17.

3.11 Erosion along the left bank due to the lateral movement of the banklines 92 in opposite direction across section D17, where the river had split into two prominent near equal sized and parallel channels. The images in the right were taken during a field visit to the site in June, 2014.

3.12 Lateral westward migration of the active channel belt from D19 to D25 93 causing erosion along the right bank and aggradation along the left.

3.13 Meander neck cut-off between D34 and D35 that shifted the channel to 94 the east

3.14 Migration between sections D24 and D28 that recorded the highest rate of 96 bankline migration during the post-dam period (2002-2014) at 36 m/y (left bank) and the maximum area of 2.5 km2 (shaded area marked as ‘e’). The figure shows how the primary channel had shifted eastward eroding along the left floodplain and reconstructing the left bankline in the east, while the right bankline persisted due to the formation of a side channel

3.15 Centerlines of the Dikrong channel for the years 2002, 2006, 2010 and 99 2014 and right banklines of the Subansiri during the same years showing the pattern of frequent shifts in the point of confluence between the two rivers

xviii

List of tables and figures

4.1 A typical chang ghar as seen in Borbil village 112

4.2 Location of RL1 (Borbil) and RL2 (Pachnoi Ujani No.1) villages along 113 the left bank of Ranganadi, (A) inset showing how RL2 is located between the Ranganadi and the Joyhing in the form of an island bar and thus is affected by flow fluctuations in both channels. The image date is March 7, 2018. (B) A second inset of RL2 dated July 27, 2006 is prior to the flash flood in 2008. It can be observed that the Joyhing although present is much smaller compared to the one in the 2018 image. The open green patches in the images represent agricultural lands (primarily under paddy) (Source: Google Earth Pro)

4.3 Images showing the long-term socio-economic effects of the flash flood in 116 2008 in RL2 village, (A) temporary settlements along the embankment, of some of the families that had been internally displaced during the flood, (B) part of the primary school in the village that was damaged by flooding waters that had breached the embankment in 2008 and 2009, (C) and (D) past and present living conditions of one of the affected and internally displaced families. The previous house was built on concrete stilts unlike the present one, and came to be completely covered by coarse sand deposition in the aftermath of the flood

4.4 Percentage distribution of households leasing in land for agriculture as 118 well as affected by sand casting in each of the surveyed villages during pre-dam and post-dam periods

4.5 Village-wise percentage distribution of households having multiple 119 livelihoods during pre-dam and post-dam periods

4.6 Embankment breach on the right bank of Ranganadi in 2002 and the 120 ensuing inundation of agricultural lands in RR5

4.7 Hydrographs of daily flow for the years 2002 and 2008, pertaining to the 120 high rainfall period of May to September. Both years were reported as impact years by the surveyed population in RR5 when long-term changes to agricultural lands had occurred, triggering changes in livelihood

4.8 Percentage distribution of households in each village according to the 136

xix

List of tables and figures

reasons cited for the downstream adversities attributed to RHEP, the sense of threat from the upstream dam and negative perception of the hydel project

5.1 Hydrographs of daily flow in Ranganadi during the high rainfall period of 158 May to September for the years 1991, 1995 and 2008. 1991 and 1995 are pre-dam years, while 2008 is post-dam

5.2 Locations of the hydroelectric projects (marked as red rectangles) planned 163 and under construction in the Dikrong basin

xx

Chapter 1

Introduction

1.1 Introduction

Hydropower has long been confirmed as a well-proven (Santos, 1991), and established technology and is generally considered a clean, carbon-free, benign and renewable source of energy (Andrews and Jelly, 2007; Ministry of Power, 1998, 2008a; Shukla, 2000; Kesharwani, 2006). The origins of generating power from water can be traced back to more than 2000 years, when the Greeks employed waterwheels to grind grain (U.S. Department of Energy, n.d.). Similar use of waterwheels (known as noria) was present in Mesopotamia and ancient Egypt to divert water from nearby streams and groundwater wells for irrigation purposes (Cech, 2010). During the Han dynasty between 202 BC and 9 AD, the Chinese used waterpower to grind grain, break ores, and make paper (International Hydropower Association, 1995-2016). It was only much later, with the inventions of electricity and turbines that hydroelectric power came to be generated.

Hydroelectricity is generated by converting the potential energy of falling water to mechanical energy which is then used to turn a hydraulic turbine connected to an ‘electric generator’ (USGS, 2016). It is renewable in the sense that the ‘fuel’ for hydropower is water, which in itself is renewable and is not consumed in the electricity generating process (Frey and Linke, 2002). One of the most important technical advantages of hydropower is its flexibility in meeting both base as well as peak electricity demands (Egré and Milewski, 2002). Hydropower schemes are robust, high efficiency, long-term investments with long useful lifetimes of usually 50-100 years and more (Oud, 2002; Ministry of Power, 2008a). A great deal of energy is lost in the transformation of primary energy into electricity, especially in case of fossil fuels. However, as per the methodology and assessment of the International Energy Agency (IEA), many renewable energy technologies such as hydropower, wind and solar, have an assumed conversion efficiency of 100%, i.e., zero conversion losses (IEA, 2015).

Since the first hydroelectric power plant constructed in Wisconsin, US, in 1882, hundreds of hydroelectric projects have come into being in later years across the world. During the early 20th century, USA and Canada were the pioneers of hydropower development, but since the last few decades have been surpassed by countries like China 1

Chapter 1

and Brazil as the ‘world leaders of hydropower development’ (International Hydropower Association, 1995-2016). China alone has over 23,000 dams ranking first in the world, followed by the United States of America with more than 9000 dams (ICOLD, n.d.; WCD, 2000). Currently, the Three Gorges Dam on the Yangtze River in China holds the record for being the largest dam in the world in terms of total installed capacity at 22,500 MW, followed by the 14000 MW Itaipu Dam on the Paraná River between Brazil and Paraguay. India, too, is not far behind when it comes to dam building, ranking third in the world with over 5000 dams (ICOLD, n.d.; WCD, 2000; Iyer, 2003).

Water harvesting, conservation, and diversion have long been practiced in India since ancient times. One of the earliest examples of water diversion and dam construction is the Grand Anikut or Kallanai Dam, built over 1500 years ago across the Kaveri River in Tamil Nadu, by the Chola Kings for irrigation purposes. Similar water management projects involving canals were constructed during the Mughal era. However, dam construction saw an escalation during the colonial rule that continued onto post- independent India when large dams were held up as the ‘temples of modern India’ by Prime Minister Jawaharlal Nehru. The first hydroelectric plant in India was a small hydropower project of 130 kW established in Darjeeling in 1897 (Bhattacharya and Jana, 2009), that marked the beginning of hydropower development in the country (NRSC- ISRO, 2015). Prior to 20th century, India had 42 large dams. The major upsurge in dam building took place between 1970 and 1989, when almost half the total number of large dams in the country were constructed (Iyer, 2003). Earlier regarded as the ‘symbols of development’, built largely for the purpose of irrigation and electricity, dam building has remained an important agenda in energy and water resources planning even today with much of the hydro potential of the country still unexploited.

As per the Reassessment study completed by the Central Electricity Authority (CEA) of India in 19871, the economically and technically feasible hydroelectric power potential of the country was estimated to be around 150 GW of probable installed capacity from a total of 845 schemes (CEA, 2008, 2017). Table 1.1 lists the basin wise

1 The Central Water and Power Commission conducted the first comprehensive survey and assessment of the hydro potential in India in 1953-1959, that estimated the economically feasible hydro potential of the country at 42100 MW at 60% load factor from 250 schemes that would generate 221 billion units of energy annually. The second assessment or Reassessment was undertaken by the Central Electricity Authority (CEA), India during the period 1978 to 1987, whereby the potential was revised to 84 GW at 60% load factor and corresponding to around 150 GW from 845 hydro schemes (CEA, 2008).

2

Chapter 1

hydroelectric potential of the country as per the 1987 reassessment studies. Despite the huge potential, less than a third of it has been tapped and only 45 GW has been installed so far (IEA, 2015). The Brahmaputra Basin accounts for the highest potential at 66 GW of probable installed capacity from 226 identified schemes, with more than 95% of it being large2 hydro above 25 MW. As of 2017 only 6% (3700 MW) is under operation and another 7% is under construction (CEA, 2017), while the rest 87% is yet to be developed. Given its vast water resources that remain largely unexploited, the Brahmaputra and Barak River basins in the Northeast, have come to be identified as the ‘future powerhouse’ of the country in recent times (Vagholikar and Das, 2010). As a result, the region is now witnessing a rapid development of hydropower projects ranging from small to large-dams in various stages of planning and construction as part of the country’s plan to expand its available domestic energy resources mix and attain energy security. Development of the hydro resources in the region had been accelerated under the 50,000 MW hydroelectric initiative3 of the Ministry of Power in 2003 (CEA, 2008; Vagholikar and Das, 2010). However, dam construction faces a lot of opposition in the Northeast due to various environmental and socio-economic impacts associated with large dam construction.

Hydropower schemes involving large reservoirs and storage dams that submerge valuable forests and lead to human displacement inevitably draw a lot of criticism (Iyer, 2003). Benefits from irrigation as well as flood control are some of the chief positive effects of dams and reservoirs, yet the actual benefits from such positive contributions may tend to vary according to the purpose of the dam and the location (Cernea, 1997).

2 In India, all projects with an installed capacity of ≤ 25 MW are termed as small hydro, while those ≥ 500 MW as mega hydro (Ministry of Power, Government of India, 2018). In this thesis, all projects with installed capacity greater than 25 MW has been defined as large hydro. The International Commission on Large Dams (ICOLD) define large dams as, “A dam with a height of 15 metres or greater from lowest foundation to crest or a dam between 5 metres and 15 metres impounding more than 3 million cubic metres” (ICOLD, 2011, p.3). 3 The 50,000 MW Hydroelectric Initiative was a programme launched by the Ministry of Power, in 2003 that involved preparation of Preliminary Feasibility Reports (PFRs) for 162 hydel projects located in 16 states across the country. The Central Electricity Authority was the nodal agency and seven consultants were hired by it to prepare the PFRs. Seventy-two PFRs were prepared for the Northeast states of Arunachal Pradesh, Assam, Sikkim, Meghalaya, Manipur, Mizoram, Nagaland and Tripura, of which forty- two were for projects in Arunachal Pradesh (CEA, 2008).

3

Chapter 1

Table 1.1: The basin-wise status of hydro-electric potential development in India in terms of the total installed capacity (as on 31.05.2017)

River Basin Identified Capacity Capacity Capacity Capacity yet capacity as per under under under to be taken assessment study operation construction operation + up for (MW) (MW) under construction construction (MW) (MW)

Total Above

(MW) 25 MW (MW)

Indus 33832 33028 13798 3358 17155 15872

Ganga 20711 20252 5317 1541 6858 13394

Central Indian 4152 3868 3147 400 3547 320 Rivers

West flowing 9430 8997 5682 100 5781 3215 Rivers

East Flowing 14511 13775 8163 1050 9213 4562 Rivers

Brahmaputra 66065 65400 3701 4284 7985 57415 Basin

All India 148701 145320 39809 10733 50541 94779

Adapted from “Status of H. E. Potential Development - Basinwise (In terms of Installed Capacity above 25 MW)”, Central Electricity Authority, December 12, 2017.

Given the complexities of the Northeast region in terms of its ecological and geological sensitivity, cultural uniqueness, ethnic diversity (Vagholikar and Das, 2010; Verghese, 2006; Hazarika, 1995), and the highly sensitive political and socio-economic issues (World Bank, 2007), the construction of large and mega dams in the Northeast,

4

Chapter 1

especially the rate at which these projects are being sanctioned has raised much alarm. Primary fears and concerns associated with multiple dams planned for the region include loss of endemic biodiversity, seismic instability, dam-induced flash floods and multiple downstream impacts and potential conflicts arising from demographic changes brought about by involuntary displacement of local population in a region already rife with ethnic conflicts.

In the state of Assam, one of the major concerns to have come up in the whole dam debate is the downstream changes in river environment and socio-economic impacts from dams in upstream Arunachal Pradesh. Assam is a floodplain zone and a downstream riparian with respect to all projects coming up in Arunachal Pradesh. All the Himalayan and sub-Himalayan rivers originating or flowing through Arunachal Pradesh pass down to the flatter floodplains of Assam, where they join the Brahmaputra River system. Assam, where floods are a persistent problem, is highly vulnerable to flow changes in the rivers and large dams in Arunachal Pradesh further increases this vulnerability by regulating and altering flows downstream.

The 405 MW Ranganadi Hydel Project (RHEP) operated by the North Eastern Electric Power Corporation Ltd. (NEEPCO), is one such hydropower scheme located in Arunachal Pradesh, which has been operating since 2002 and involves transfer of water from the Ranganadi River to an adjacent river – Dikrong, for power generation. The primary aim of this study is to analyze the impact of this large hydel project and the associated water diversion, on the riverine environment (specifically the hydrology and channel morphology) of the two affected rivers in their downstream reaches within the floodplain zone of Assam. Therefore, the study concentrates on the ‘off-site’4 impacts of the project. The study also examines how the physical changes in the Ranganadi following dam operations have affected upon the socio-economic conditions and riparian dependence of the downstream floodplain communities in the basin. The first two types

4 Slootweg et al. (2001) in their study ‘Function evaluation as a framework for the integration of social and environmental impact assessment’ discuss the concept of ‘on-site’ and ‘off-site’ impacts of biophysical changes caused by human intervention of a natural resource. They explain that, while most anthropogenic interventions usually result in impacts in the immediate area of the alteration termed as ‘on-site’ impacts, some biophysical changes impact over a wider geographical range away from the site of intervention, thus termed as ‘off-site’ impacts. Hence, this study concentrates on such off-site impacts of dam development on the downstream riparian plains of Assam approximately 30 km away from the Ranganadi hydel project.

5

Chapter 1

of impacts (i.e. hydrological and channel morphology) fall under the broad category of environmental impacts experienced by rivers following dam interventions.

1.2 Study area

The Ranganadi hydel project encompasses the adjacent basins of Ranganadi and Dikrong Rivers, which are north-bank tributaries of the Brahmaputra River system in the country’s Northeast region (Fig. 1.1). As a hydroelectric project with inter-basin water diversion, the dam and reservoir of RHEP is located over River Ranganadi at 27°20/ N latitude and 93°49/ E longitude, at Yazali in the Lower Subansiri district of Arunachal Pradesh. The powerhouse of the project, on the other hand, is located on River Dikrong in the adjacent basin, at Hoj in Papum Pare district, Arunachal Pradesh. Assam forms the lower riparian state and encompasses the alluvial floodplain zone of the two rivers, through which they traverse to finally confluence with River Subansiri, which joins the main-stem Brahmaputra further downstream. Within Assam, the downstream catchment falls in the Lakhimpur District, a primarily agrarian area.

Planning for the project began in 1978 and was finally assigned to NEEPCO for preliminary investigations and later construction and operation (Das and Ahmed, 2005). Project construction started in 1988 and fully commissioned on March 27, 2002. The RHEP involves a concrete-gravity diversion dam of 68 m height and 344.75 m length that diverts 160 m3/s of water through a 10 km tunnel from Ranganadi to Dikrong. With a total installed capacity of 405 MW, the project has three 135 MW capacity turbines to generate electricity. The current project is Stage I of the two projects planned on the Ranganadi. Stage II with an installed capacity of 130 MW and a rock-fill embankment dam has been designed upstream to store water for Stage I. Power generated is distributed amongst all the seven states of the Northeast – Assam, Arunachal Pradesh, Meghalaya, Mizoram, Tripura, Nagaland and Manipur. Table 1.2 contains some of the technical features of the project.

Like all other north bank tributaries of the Brahmaputra, steep gradient, heavy discharge and sediment loads, and high rainfall regimes characterize the Ranganadi and Dikrong rivers. Both the rivers pass through a high gradient, rugged hilly terrain in Arunachal Pradesh while opening onto flatter floodplains in Assam. Due to the dynamic

6

Chapter 1

Fig. 1.1 Study area showing the two concerned rivers - Ranganadi (impounded river) and Dikrong (flow-recipient river), (a) location of the dam and reservoir in the former, (b) location of the project’s powerhouse in the latter, (c & d) starting points of the study lengths in each river. nature of the rivers, their riparian zones are prone to frequent flooding with considerable impacts on the socio-economic conditions of the riparian populations. The vast floodplains, especially the river banks, are predominantly utilized for agricultural activities (agriculture being the primary economic activity) that include cultivation of paddy as the major crop. Besides summer and winter rice, rabi crops are cultivated during the winter season with the river banks providing adequate fertile space for these crops to be grown. The rivers support many wetlands by their banks that not only have great ecological significance in the floodplain landscape but also provide ecological goods and services to the riparian populations. These wetlands are an important and rich source of fish protein. Fishing is a regular riparian activity for non-commercial consumptive as well as recreational purposes.

7

Chapter 1

1.3 Literature review

1.3.1 Energy scenario in India and the rationale behind hydropower development

India, today, faces a highly challenging task of meeting the ever-increasing demands for energy amidst growing urbanization and industrialization. It is currently the third largest economy in the world with a total population of 1.3 billion that accounts for approximately 18% of the world’s population. The United Nations has projected that India will surpass China as the most populous state in 2025/2050 reaching a total population of more than 1.5 billion. Given its rapidly expanding economy, a growing population and rapid industrialization and urbanization, the energy demand by all sectors has also increased sharply. Energy use of the country in 2011 was the fifth highest in the world having increased by 16 times over the last 60 years (Garg, 2012). As per the 2015 India Energy Outlook Report by the IEA, energy use in India is projected to more than double by 2040 as it reaches 1900 million tons of oil equivalent. Accordingly, the country needs to increase its primary energy supply by 3 to 4 times, its electricity generation capacity by 5 to 6 times and the crude oil requirement by 4 times of its 2003-04 levels, if it is to maintain a steady growth rate of 8 to 10% over the next 25 years (Planning Commission, 2006; Garg, 2012). Thus, by 2031-32, the power generation capacity has to increase from the current capacity of 160 GW to 800 GW.

In order to meet the growing energy requirements in a technically efficient, economically viable and environmentally sustainable manner (Planning Commission, 2006), an instrumental step in the country’s energy road map was to expand its energy resource base and tap into all possible domestic sources of energy. According to the Planning Commission’s report (2006), this involves increased emphasis on renewables and increasing their share in the energy mix. The report further confirms that renewables had accounted for about 32% of primary energy consumption in 2003-2004, with traditional biomass used for cooking and electricity from large hydro being the major contributors. However, the energy situation in India can be summarized as one where there is an ever-increasing demand for energy with little reduction in fossil fuel dependency, while a large portion of the population still do not have access to modern energy services (Nathan et al., 2013).

8

Chapter 1

Table 1.2: Features of the Ranganadi Hydel Project Total Installed Capacity 405 MW Turbines 3 × 135 (Francis turbines) Design Energy 1509.66 MU HYDROLOGY Catchment area up to Yazali dam site 1,73,000 ha Catchment area up to diversion dam site 1894 sq.km Annual Rainfall Maximum 1562 mm Minimum 838 mm Average 1232 mm Reservoir Characteristics Maximum annual run-off 5080 m3/s Minimum annual run-off 2048 m3/s Maximum water level 567 m Full reservoir level (FRL) 567 m Water spread area at FRL 160 ha Diversion Dam and Tunnel Dam type Concrete gravity Dam height 68 m Dam length 344.75 m Spillway capacity 9175 m3/s River bed level 523.5 m Foundation level 505 m Tunnel diameter 6.8 m Tunnel length 10.13 km Tunnel capacity 160 m3/s

Adapted from “REPORT ON MONITORING THE IMPLEMENTATION OF ENVIRONMENTAL SAFEGUARDS OF RANGANADI HYDRO ELECTRIC PLANT, Period from 1st April’2012 to 30th September, 2012,” by NEEPCO, 2012, pp. 1-12, and “Ranganadi Hydroelectric Project: Features”, NEEPCO, 2017. Copyright 2017 by North Eastern Electric Power Corporation Ltd.

9

Chapter 1

Electricity is one of the prime drivers of economic development (Sridharan, 2009) and one of the basic energy services entitled to people. The industry and buildings sector are two of the major consumers of electricity in India and within the buildings sector, the rising ownership of electrical appliances in both rural and urban areas have pushed the electricity demands higher (IEA, 2015). During 2000 to 2013, electricity demand by the buildings sector increased at an average annual rate of 8%. As per the IEA report (2015), overall electricity demand of the country increased from 376 TWh in 2000 to 897 TWh in 2013, with an average annual increase rate of 6.9%. The power sector currently constitutes around 15% of the final energy consumption and is projected to grow further and faster.

On the other hand, the Indian electricity scenario suffers from high demand-supply gap that has only widened over the years and despite an increase in generation capacity along with increased and improved power availability, demand has ‘consistently outstripped supply’ (CEA, 2012; Garg, 2012). According to the CEA’s National Electricity Plan (2012), peak and energy shortages still prevail due to inadequate generation and high transmission and distribution losses of around 29.24%. India is the third largest producer of electricity in the world along with an overall increase in access to electricity for all households from 55.8% in 2001 to 67.2% in 2011, yet 44.7% of rural households and 7.3% of urban households still remained un-electrified as of 2011 (Census of India, 2011). Even for households with electricity, connections suffer from low loads and frequent power cuts (Saxena and Kumar, 2010). Per-capita electricity consumption5 reached 1010 kWh in 2014-2015, yet it is quite low compared to countries like China, which averages at 4000 kWh, while for developed nations, per-capita consumption is 15,000 kWh (CEA, 2016, 2017; Bhaskar, 2015).

Current power generation capacity of the country is about 290 GW (IEA, 2015) from a diverse energy mix; but conventional energy sources that are also major greenhouse gas emitters make up the largest share (Sen et.al., 2016). As on 31st March, 2017, the total installed capacity of the country was 327 GW, of which 218 GW (67%) is thermal based (coal, gas and diesel), 44 GW (13%) is hydro (> 25 MW), 6 GW (2%) is

5 Per-capita consumption of electricity = (Gross generation + Net import)/ Mid-year population (CEA, 2017).

10

Chapter 1 nuclear and 57 GW (17%) is renewable energy sources (RES)6 (CEA, 2017). Accordingly, total electricity generation from all sources during April 2016 to February

52.57 3910.19 Total Grand 89581.59 91808.36 34406.87 107088.96 326848.53

Central Electricity Electricity Central

12.52 990.74 281.12 RES* 11539.36 18304.43 26132.07 57260.23 (MNRE)

0.00

7447.50 4738.12 1242.00 Hydro 19311.77 11739.03 44478.42

up - 0.00 0.00 0.00

1620. 00 1840.00 3320.00 6780.00 Nuclear

wise break -

40.05 Mode 2387.07 Total 79497.03 60617.26 28678.02 57110.46 218329.88

0.00 0.00 0.00

36.00 40.05 837.63 761.58 Diesel

Thermal 0.00

Gas 100.00 5781.26 6473.66 1771.05 11203.41 25329.38

DIA INSTALLED CAPACITY (IN MW) OF POWER STATIONS”, MW) POWER (IN OF CAPACITY INSTALLED DIA

0.00 . 2017 580.02 Coal 51329.20 6 8293.62 43382.02 28578.02 192162.88

North Islands Eastern Region Eastern Western All India Northern Southern *RES = renewable energy sources. sources. energy renewable = *RES All India Installed Capacity of Power Stations (Utilities, as on 31.03.2017) (in MW) (in on 31.03.2017) as (Utilities, Stations Power of Capacity Installed India All Table 1.3: IN “ALL from Adapted 31, , March Authority

6 Large hydro is also a renewable source of energy but is no longer considered in the new and renewable energy mix. Large hydro comes under the regulation of the Ministry of Power while Small Hydro Power (SHP) is governed by the Ministry of New and Renewable Energy (MNRE). The modern renewable energy sources (RES) detailed in Table 1.3, includes Small hydro power (SHP), Biomass Power (BP), Urban and Industrial waste power (U&I), solar and wind power (Tripathi et al, 2016; CEA, 2017).

11

Chapter 1

2017 was 1137 TWh, of which 1061 TWh was from conventional sources and the rest 76 TWh from renewable sources (CEA, 2017). Renewables have increasingly become more important in Indian energy planning in a bid to reduce the adverse environmental impacts of thermal based power generation. Within the conventional sources, thermal based generation was highest at 86% followed by hydro that generated 11% of the total electricity during the above-mentioned time-period. Table 1.3 shows the installed capacity statistics in India at the end of March 2017.

Thus, it can be observed that hydropower constitutes the second largest source of electricity in the country, while a huge potential of another ~ 90 GW corresponding to 65% of the total assessed large hydro, still, remains to be harnessed. As of May 2017, the total capacity of hydropower under operation and construction constituted 50 GW, which is only 35% of the assessed potential (CEA, 2017) (Table 1.1). Therefore, much is still needed to be accomplished with respect to the second largest contributor to electricity generation in the country. The CEA called for large-scale mobilization of the untapped hydro potential of the country to ensure assured domestic supplies of power and meet future electricity demands. Under the National Electricity Policy (2005), the Ministry of Power, also initiated several reform measures for the addition of new generation capacity, under which once again, maximum emphasis was laid upon full development of the feasible hydro potential of the country. As a result, the Northeast region of India, which forms a major part of the Brahmaputra basin having the highest yet largely untapped hydro potential, is now witnessing a rapid construction of hydropower projects that include some very controversial large dams.

1.3.2 Northeast hydropower scenario

According to the proponents of hydropower projects, development of the Northeast hydro potential would contribute both towards the region’s economic progress as well as towards power supply for the rest of the country. Advocates of dams also point towards the need for flood control in the region, specially, in the downstream floodplain areas that can be realized through dam construction. Moreover, hydro can play a significant role in improving the internal power situation within the region. The Ministry of Power in its National Electricity Policy (2005) stated that the government should emphasize speedier harnessing of the country’s untapped hydro potential and regions like the Northeast should focus on maximum development of hydel projects, since it would facilitate the

12

Chapter 1

economic development of the region. However, despite the likely benefits to be gained, realizing the same in a swift manner has proved to be difficult and is marred with obstructions and cost overlays. Rao (2006) highlights the technical issues involved in transmission and evacuation of the generated power to the load centers, owing to the lack of developed and accessible routes to the remote dam sites. Additionally, there needs to be cooperation at the local level, whereby the environmental and social issues related to dams with large storage schemes are adequately addressed and the uneven distribution of costs and benefits are minimized (Rao, 2006).

The World Bank (2007) in its Strategy Report titled, ‘Development and Growth in Northeast India: The Natural Resources, Water and Environment Nexus’, put forward similar insights. The report pointed to the revenue that the region would earn from the export of hydroelectricity to the power starved Northern and Western states of India. It also highlighted the other potential benefits such as flood moderation and reduced flood damages in Assam from storage schemes in upstream Arunachal Pradesh, the generation of employment opportunities and secondary positive impacts on services, transport, and tourism. At the same time, the report also emphasizes addressing adequately the cultural and territorial identity of the tribal communities in the region and their livelihood security.

Furthermore, the World Bank study calls for an integrated analysis of the trade- offs of different development options that would fully take into account the diverse interests of the Northeastern Region and its citizens. One of the strategies for sustainable development of the hydro potential in the region, suggested by the study, is to assess the possibility for identifying and developing certain low-impact projects first that would allow some progress to take place in the socio-economic front while agreement is reached on those sites with larger storage potential. This would create scope for development while not completely ruling out storage schemes facing strong opposition on various environmental and social grounds. However, the study still fails to capture the concerns of downstream populations.

In addition to the extensively studied upstream impacts of dams, large projects such as those in Arunachal Pradesh also result in widespread impacts downstream of the dam. In fact, it is the cumulative downstream impacts of dams in Arunachal Pradesh, which has become a major cause of concern and opposition to hydropower development

13

Chapter 1

in the downstream riparian state of Assam. Livelihoods of most of the downstream communities in the largely agrarian state is natural resource based and intricately linked with floodplain ecology, that is likely to be affected by irregularities in the natural flow regime of the rivers (Vagholikar and Das, 2010).

1.3.3 Downstream impacts of dams

Petts (1987) categorized three orders of downstream impacts that occur in rivers following dam closure. First-order changes are the direct changes and disruptions in downstream sediment and water flow including changes in the magnitude and duration of flows, alteration of the flood frequencies and timing, changes in water quality, planktons etc. The second-order changes primarily result from the first-order hydrological alterations and include the changes in channel morphology such as channel width, depth, channel pattern etc. Third-order changes result from a combination of the first two and mainly include the alterations in riverine ecology, fauna etc. Thus, “by changing the flow of water, sediment, nutrients, energy, and biota, dams interrupt and alter most of a river's important ecological processes” (Ligon et al., 1995, p.183).

Similar categorization of the ecological impacts of dams was also put forward in the World Commission on Dams (WCD) report in 2000, where the hydrologic and geomorphic changes in a river were classified as first-order impacts. Second-order impacts involved the effects and changes on primary biological productivity of riverine ecosystems such as effects on riparian plant-life and wetland habitats. Third-order impacts involved alterations to fauna (such as fish) caused by a first-order or second- order effect on the river. The WCD report is considered one of the most comprehensive contributions to global dam literature. The Commission did an extensive and rigorous global review of large dams that included eight detailed case studies of large dams, country reviews for India and China, a briefing paper for Russia and the Newly Independent states, a crosscheck survey of 125 existing dams, 17 Thematic Review papers, and large number of consultations. However, with respect to downstream impacts, even such an extensive review exercise by the WCD failed to capture the extent and magnitude of downstream damages of dams.

As stated by Richter et al. (2010, p.15), the WCD report did succeed to bring “much-needed global visibility and media attention to the benefits and costs of large

14

Chapter 1

dams”, but failed to give due recognition to the plight of downstream populations whose livelihoods also happen to be affected by hydrological alterations in a river downstream of the dam. It concentrated more on the social impacts and challenges of resettlement and rehabilitation pertaining to populations living in the upstream reaches of a dam.

1.3.4 Impacts on downstream river hydrology and geomorphology

The natural flow regime of a river is the most important component that maintains the ‘ecological integrity’ of the river and riparian ecosystem (Poff et al., 1997). The magnitude, duration, timing, frequency, and rate of change of the hydrologic conditions are the five critical components of flow regime that regulate the ecological processes in a riverine system and ensure exchange of water and sediments with its riparian areas (Petts, 1985; Poff et al., 1997; Richter et al., 1996, 1997). Dams, irrespective of whether they are built for irrigation, flood control, or hydroelectricity purposes, significantly alter these critical components of natural flow in the downstream reaches of a river.

The primary effects of dams on downstream hydrology are through changes in the high flow and low flow conditions whereby high flows are reduced and low flows are typically increased (Magilligan and Nislow, 2005). Graf (2006) reported 67% reduction in annual peak discharges, 60% decrease in the annual maximum/mean flows, 64% decrease in the range of daily discharges, 34% increase in the number of flow reversals along with changes in the timing of yearly high and low flows by almost half a year, in 72 dam regulated river reaches across America. Similarly, Magilligan and Nislow (2005) measured the hydrological alterations following impoundment at 21 gage stations across the United States and observed significant increase in the 1-day to 90-day minimum flows and ~39% to ~55% decline in 1-day through 7-day maximum flows.

Flow reductions were reported from the Correntes River in Brazil due to the hydroelectric operations of the Ponte de Pedra reservoir but were considered modest alterations compared to dam effects on other rivers. According to Fantin-Cruz et al. (2015), this could be attributed to the ‘small regulating capacity of the reservoir’ and the naturally existing seasonality in precipitation and run-off patterns. The cascade of dams upon the trans-boundary Lancang-Mekong River in China and Lower Mekong riparian countries have caused significant decline in downstream water discharges during both the flood and dry seasons along with high fluctuations and decrease in the yearly sediment

15

Chapter 1

flux (Lu and Siew, 2006; Fan et al., 2015). This has resulted in negative impact upon downstream riverine flora and fauna, specially the fish assemblages.

The hydrologic changes caused by dams result in significant modifications in downstream river geomorphology. Channel morphology encompassing the dimensions of channel geometry such as length, width, depth and slope, channel types such as braided, sinuous, or meandering, is primarily a function of discharge and sediment load (Galay, 1983; Williams and Wolman, 1984; Rosgen, 1994; Petts and Gurnell, 2005; Lord et al., 2009). However, the channel responses to upstream impoundment and flow regulations can be quite complex and vary according to the basin and river type, topography, geology, channel bed and bank materials etc. (Williams and Wolman, 1984; Shields Jr. et al., 2000). Channel narrowing is the most frequently observed geomorphic effect of flow reduction by dams (especially reduction in peak flows or bankfull discharges). The effect has been reported in many rivers such as the Bill Williams River (Alamo dam, USA), the Peace River (Bennett Dam, Canada), and the Piave River (multiple dams and barrages, Italy) (Church, 1995; Surian, 1999; Shafroth et al., 2002). Zahar et al. (2008) reported reduced channel capacity, channel narrowing, and aggradation of the riverbed of the Medjerda River in Tunisia following flow reductions by the Sidi Salem dam. The reductions in channel capacity have alternately lead to occurrences of large floods.

The hydrological alterations in the regulated rivers reported by Graf (2006, p.336) resulted in the dam affected reaches to have 37% lesser geomorphic complexity as compared to similar unregulated reaches and also reduced size of the ‘hydrologically active functional surfaces’. Likewise, reductions in downstream flood magnitudes and efficient sediment trapping by the reservoir of the Windamere Dam in Australia resulted in ‘complex channel responses’ in the river. This included simultaneous channel narrowing and morphological adjustments, degradation, and aggradation (with increase in the bed-material size), increased formation of bars at the tributary mouth, and ultimately ecological changes such as large-scale vegetation establishment at various downstream river sites (Benn and Erskine, 1994).

1.3.5 Downstream socio-economic impacts

In addition to the three orders of changes discussed above, there is a fourth order of downstream impacts that does not stay limited to the biological and physical environment

16

Chapter 1 of the river, but also extends to impact upon the socio-economic conditions of the riparian populations living alongside the river. The downstream socio-economic impacts mostly result from impacts upon the hydrological and biophysical characteristics of the riverine/riparian ecosystem (Richter et al., 2010; WCD, 2000; Beck et.al; 2012) or broadly the downstream environmental impacts. ‘Downstream changes in agro- production systems’ (Cernea, 1997) are one of most widely documented socio-economic impact of dams in the downstream riparian areas. However, most often these relationships are poorly documented and overlooked during dam development plans. This is especially true in the Indian context. The issue is that the degree to which dams affect natural food productivity of river ecosystems that could subsequently lead to disruption of livelihoods and cultures dependent on these ecosystems is still not accurately accounted.

Floodplains are an important feature of riverine ecosystem whose integrity is dependent on the natural regimes of flooding and inundation (Richter et al., 2010; WCD, 2000; Ligon et al., 1995). These floodplains are fertile grounds used extensively for agriculture, livestock grazing, fishing, etc., and are hence an important source of livelihood for many riparian communities across the world. In his book, ‘The Future of Large Dams’, Scudder (2005) presents a detailed description of downstream floodplain ecosystems and the innumerable river-basin communities whose livelihoods are ‘dependent on free-flowing rivers or were so dependent before dam construction’.

Flood recession agriculture where cropping takes place after the flood waters recede and a succession of crops are cultivated utilizing the locked moisture in the damp soil is one of the most widely practiced cropping system in the floodplains (Scudder, 2005; Richter et al., 2010). Most of the river basins support large-scale recession agriculture or wetlands agriculture based on the natural regimes of recurring floods, which enrich the soil with rich nutrients (Adams, 1985; Cernea, 1997; Scudder, 2005). Some famous examples are the Amazon basin in Latin America, the Ganges- Brahmaputra-Meghna basin in India and Bangladesh, the Zambezi, Zaire, Senegal, and Niger basins in Africa, and the Mekong and Yangtze basins in South-east Asia.

Cernea (1997) discusses how the annual flooding patterns of these rivers are important to the specific agricultural and cropping patterns of the riparian communities that have evolved through centuries of adaptation and harmonization with nature. The associated culture is an ideal example of social and environmental integration whereby

17

Chapter 1

the local cultivators have been able to incorporate the recurring floods into their agricultural strategies. Examples could be cited of traditional rice varieties such as the African rice species (Oryza glaberrima) and Sorghum (Sorghum bicolor), that are tolerant to prolonged water submergence and floods and is cropped in most of the African river basins. However, the hydrological alterations caused by dams (specially the changes in the flood characteristics) often result in decreased downstream agricultural productivity and subsequent impoverishments. Thus, centuries of creative human adaptation to natural opportunities have been rendered obsolete and impossible (Cernea, 1997).

In the Sokoto river-valley of Northwest Nigeria, around 50,000 river-dependent people reside who utilize the natural resources of the floodplain, and almost 90% of the floodplain is put to agricultural purposes (Adams, 1985). Cropping is distributed across the upland rain-fed (tudu) land and the seasonally inundated fadama land. Fadama cropping is a complex but well-adapted agricultural system that has been attenuated to regular wet season flooding. It involves cultivation of flood tolerant crops such as rice and sorghum during the wet season, followed by cotton, groundnut, and peas in the drier floodplain areas and various vegetables in the damper areas during the dry season that utilizes the residual moisture of the soil. However, as reported by Adams (1985), the Bakolori dam constructed in the mid-1970s, reduced the magnitude, duration and depth of wet season flooding which in turn resulted in a shift from rice to lower value millet and sorghum cultivation in the fadama land. There was significant decline in fishing as well during both the wet and dry seasons and in some downstream villages, fishing ceased to exist along with out-migration of villagers.

Similar to the African floodplain agricultural systems, the floodplains of Assam are also extensively used for cultivation purposes, with rice being the major crop. Most of the rivers flowing into Assam from Arunachal Pradesh, such as the waters of the Subansiri River, in the lower reaches recharge wetlands that are in turn used by local communities such as the Mishing tribe for ‘deep water rice’ (Bao) cultivation and fisheries (Vagholikar and Ahmed, 2003). Bao is a traditional floating rice variety that can withstand water depths of greater than 100 cm and sown (broadcasted) at the onset of monsoon, i.e., April/ May and harvested in Dec/Jan (Ahmed et al., 2011). The riverine islands and tracts, locally known as chapories, are also accessed for grazing cattle and rearing livestock for dairy-based livelihoods (Vagholikar and Das, 2010).

18

Chapter 1

Besides agriculture, fisheries are another example of the myriad and complex relationships between river-dependent flood-plain communities and the riverine ecosystem. Floodplains provide fertile grounds for a variety of fish species for spawning, feeding and rearing on them. Dams, by obstructing the natural annual flooding patterns of a river, disrupt the lateral exchange of sediments and other essential nutrients between the river and riparian ecosystem, thus affecting the fish populations. Though measures have been taken in dam development plans to address the loss of fish diversity, productivity and movement to some extent, yet these plans do not adequately address the loss incurred by downstream fishing communities.

Thus, given the extensive floodplain economies flourishing across the world, the ecological and economic impacts of dams would be much more serious and higher in the downstream areas (Scudder, 2005). According to Scudder (2005), the impact would be higher, especially with respect to the discussed tropical and sub-tropical river basins that have higher biological diversity and productivity and majority of the river-dependent downstream populations are rural agrarian and fishing communities with low-incomes. At the same time, dam regulations result in a simplified and shrunken downstream river morphology which in turn changes the riparian ecology that becomes less diverse and ‘spatially smaller’ (Graf, 2006). Hence, dam constructions would further impoverish these people, yet the issue fails to gain necessary recognition in impact assessment studies.

One of the most comprehensive and well-proven studies on downstream impacts of dams has been by Richter et.al (1996, 2010). First in 1996, they proposed a methodology for assessing the hydrologic alterations within ecosystems, whereby 32 parameters referred to as ‘Indicators of Hydrologic Alteration’ (IHA) were introduced. The IHA model has also been used in this thesis to evaluate the post-dam flow changes in the Ranganadi and Dikrong Rivers. In 2010, Richter et al., gave the first global estimate of the number of river-dependent people potentially affected by dam-induced changes in river flows and other ecosystem conditions at 472 million. The study, moreover, delineated a methodological structure for identifying the potential affected downstream populations. Citing their estimate as a conservative one, the identification of the potentially affected people was based upon – a) their proximity to dammed rivers (i.e. within 10 km radius), b) the exposure of the landscape (i.e. river deltas and slopes <1

19

Chapter 1 degree are presumed to sustain fisheries and flood-plain agriculture), and c) the degree of flow regulation that a river experiences. The same factors were also taken into consideration while selecting the sample villages along the Ranganadi and Dikrong Rivers in this thesis.

Finally, Richter et al. (2010, p.31), stated that, “globally around 232 million people live in rural areas downstream of large dams close to (< 10 km) very large impacted rivers (discharge > 1000 m3/s) for which the total upstream storage capacity exceeds 10% of the annual flow”. Most of these populations reside in Southeast Asia and India.

There has been no comprehensive post-dam study on the Ranganadi Hydel project, which is the subject of this thesis, that looks at the entire range of probable impacts, i.e., the hydrologic, geomorphic, and socio-economic aspects, along with their complex inter-linkages. Rampini (2016) is one of the very few works that has studied the various socio-economic problems faced by the downstream populations from the riverine changes caused by the Ranganadi project. However, the study lacks analysis of the hydrological alterations and the resultant changes in downstream channel morphology and processes. The study provides an understanding on how hydropower development of the Brahmaputra River system increases the vulnerability of downstream riparian communities to the impacts of climate change.

1.4 Statement of the problem

For a complete assessment of impacts and accounting of costs and benefits of large scale hydropower projects such as those in Northeast India that could both economically benefit the region as well as lead to loss of livelihoods alternately, there needs to be an assessment of the various dependencies of the downstream communities on the riverine ecosystem. However, if we look at the Indian hydropower scenario and the related impact assessment studies, it is evident that there are serious limitations. Socio-economic impact evaluation remains primarily restricted to the upstream areas. Thus, only those damages arising out of reservoir induced submergence and resultant direct displacement of local and indigenous upstream populations are considered as project-affected-people (PAP). The associated Resettlement and Rehabilitation policies of the country are also structured to satisfy the specific needs of only the upstream PAPs while downstream affected

20

Chapter 1

populations are completely kept out of consideration. In addition, while downstream environmental impacts such as impact on the hydrology of the river, effects on the fish populations, etc., are analyzed in the Environmental Impact Assessment (EIA) process yet these do not extend to the identification and analysis of the associated livelihood dependence of downstream communities on these ecosystem goods and services.

This might lead to a serious underestimation of the number of people actually affected by large dams and calls for urgent review. Specially, in the context of the Northeast region, which is currently witnessing a rapid construction and sanctioning of large dams, realizing the downstream socio-economic impacts necessitates top priority. Inadequate attention towards such impacts is a major cause of discontent and resentment amongst the local communities towards government agencies and proponents of dams in downstream Assam.

1.5 Why the Ranganadi hydel project?

Firstly, RHEP is the only operating large dam or hydroelectric project in Arunachal Pradesh, where majority of the future large dams are being planned and constructed. A second key feature of RHEP is the water diversion aspect, which is again the only type present in the region, and more so involving rivers with similar geographic, climatic, hydrologic and geomorphic characteristics, being located in adjacent basins. The resultant impact has changed the fluvial characteristics of the two rivers in opposite patterns, the detailed analysis of which has been presented in this thesis. This is important because although it has been 15 years since the project started operating in 2002, there has been no analysis of the extent of post-dam hydrological alterations caused by the project on the downstream flow conditions of the two affected rivers. The persistent occurrence of floods, especially flash floods, in downstream Assam is an added concern that further necessitates examining the impact of the dam upon the downstream flood regime of the rivers for better flood control planning.

The current project is not the only project planned over the Ranganadi and Dikrong Rivers. Stage II of the Ranganadi Hydel project is yet to be constructed upstream of the current location at Yazali, while Dikrong River is set to have three more projects of which one would be upstream of the powerhouse at Hoj, and the other two downstream of it. Given the cascade of projects planned over the two rivers, it is even more important to

21

Chapter 1

study the on-going impacts of the existing project in order to have better management strategies at place.

Therefore, the primary research question that arises is ‘how has the Ranganadi hydel project impacted the downstream hydrological and geomorphological characteristics of the Ranganadi and Dikrong Rivers, given the flow diversion, and how have these physical changes in the riverine system impacted upon the socio-economic conditions of the riparian populations in the adjoining floodplain of the impounded river?’. The question being raised in this thesis is situated within the larger debate on large dam construction in Northeast India, specially Arunachal Pradesh, and the potential impact of such projects on the downstream reaches of the affected rivers in the fertile, yet highly vulnerable floodplain areas of Assam.

1.6 Objectives

(1) To assess the downstream environmental and associated socio-economic impacts of the Ranganadi hydel project.

(i) To analyze the downstream temporal and spatial hydrological and geomorphological changes of Ranganadi and Dikrong Rivers.

(ii) To identify the socio-economic structures of downstream riparian populations and their river dependence.

(iii) To analyze the downstream socio-economic impacts post-dam construction.

(2) To analyze the sequence and pathways by which the downstream environmental effects in the river finally impact upon the socio-economic structure of riparian populations.

(3) To analyze and identify the gaps and limitations in existing programmes and policies with respect to the study of downstream impact assessment in India.

1.7 Chapter outline

The thesis has been divided into six chapters. The first chapter is introductory and outlines the topic of research, the statement of problem and primary research question,

22

Chapter 1

and the associated literature review. It states the primary research objectives, adopted methodology, data sources and related research materials.

Chapter 2 discusses the hydrological alterations in the downstream flow regime of the Ranganadi River, following dam obstruction and water diversion in 2002. This chapter also defines the morphological features of the Ranganadi channel such as channel braiding and sinuosity, width, bankline migration, erosion and deposition and their modification across time and space. Correlation analyses between flow characteristics and channel changes determine how the dam has influenced the changes during the post-dam period.

Chapter 3 is similar to the second chapter and presents the hydrological alterations in the flow-recipient river, i.e., the Dikrong River, following addition of water from Ranganadi since 2002. Temporal and spatial changes in channel pattern have been defined and analyzed by comparing between the pre-dam and post-dam periods. Correlation analyses between the downstream flow pattern and channel characteristics define whether the temporal changes in the river could be attributed to the dam operations.

Chapter 4 discusses the impacts of impoundment and flow diversion of the Ranganadi River on the social and economic conditions of the downstream riparian communities in the lower floodplains of the river. River-use and dependence of the communities upon the ecosystem goods and services, their livelihood structure and how (or whether) these have been affected by dam-related changes in the Ranganadi have been discussed. This chapter also includes the different perceptions of the downstream population towards the hydel project and their observations of the post-dam changes in the riparian ecosystem.

Chapter 5 details how the two affected rivers – Ranganadi and Dikrong, exhibit opposite trends of change especially after the dam came into being. The order and pattern in which the hydrological and physical changes in the Ranganadi have affected the socio- economic conditions of the downstream communities along the river are discussed. This chapter also argues how the complex overlapping of multiple natural and anthropogenic factors has aggravated the downstream consequences of river damming and water diversion. Finally, the chapter discusses the on-going process of hydropower development

23

Chapter 1 in Arunachal Pradesh and the inherent issues related to such dam building in the Himalayan riparian, including the policy issues.

Chapter 6 summarizes the findings of the thesis and the interpretations and conclusions drawn therein with respect to dam-induced environmental and socio- economic impacts of large dams on the downstream floodplain zones of river systems, using the case study of the Ranganadi Hydroelectric project. It highlights the mitigation strategies that could be strengthened further in the Northeast hydropower scenario and stresses upon the importance of studying projects that are already functioning, to enable better adaptation strategies for projects that are under construction or in the process of being undertaken.

24

Chapter 2

Hydrological and Morphological Changes in River Ranganadi

2.1 Introduction

Rivers and their floodplains constitute a highly complex interconnected system vulnerable to frequent disturbances and changes in their geomorphological, hydrological, and ecological environment (Bryant and Gilvear, 1999). Changes in the geomorphology and fluvial dynamics of a river are not exclusive artificially induced processes but a natural progression of the river to attain a state of equilibrium. However, anthropogenic interventions like river engineering can sometimes act as the catalyst in the natural changing process and lead to its acceleration or modify the natural rate of change. As Rosgen (1994, p.181) had stated, “the rate and direction of channel adjustment is a function of the nature and magnitude of the change and the stream type involved” and while some river systems may change rapidly, others may take a long time in responding to changes – natural or anthropogenic. The simplest of the geomorphological adjustments that rivers often undergo post-dam closure and especially downstream of dams, lead to significant and at times ‘disastrous’ alterations in their ecological functioning (Ligon et al., 1995). Dams tend to change the fluvial dynamics of a river through dam associated flow regulations for power generation and amplify the probability of riverine hazards. For example, the lateral movement of alluvial rivers through erosion and deposition of banks greatly affect floodplain land-use by riparian communities (Shields Jr. et al., 2000). Therefore, it is important to study the modifications that occur in the riparian environment post-dam closure in order to understand both their short-term and long-term environmental and socio-economic impacts.

In this chapter, I discuss the wide range of hydrological and geomorphological changes undergone by the Ranganadi River in its downstream length, especially within the floodplain zone, following impoundment by the upstream hydel project. The Ranganadi constitutes the dammed or the flow-deprived river in the RHEP. Planform features of the river that were specifically examined for dam-induced changes are channel width, sinuosity and braiding intensity and the processes of bankline migration, bank erosion and deposition.

25

Chapter 2

2.2 Study area - Ranganadi basin

The Ranganadi is a north-bank tributary of the Brahmaputra with its origin in the Dafla hills in Arunachal Pradesh, at an altitude of 3440 m. The basin is bounded between 27°0/ to 27°45/ N and 93°15/ to 94°10/ E, comprising a catchment area of 2940 sq.km, of which 76.2% is in the hills of Arunachal Pradesh and 23.8% is in the plains of Assam (Agarwal et al., 2007). Thus, the river is trans-boundary, with Arunachal Pradesh as the upper riparian (where it is known as the Panyor) and Assam as the lower riparian. Throughout its course from the origin, the river receives a number of tributaries in the hills. Up till the diversion dam site (i.e. the RHEP dam), the river receives the Niyorke Nala (at 1263 m altitude), the Pering, the Pak, the Kale and the Pit Nala (Das and Ahmed, 2005). Downstream from the dam, the Ranganadi takes an eastward turn and finally flows down to the south (before reaching Kimin, Arunachal Pradesh) into the plains of Assam. It enters Assam approximately 6 km upstream of Joyhing Tea Estate (Lakhimpur district) at 27°1/ N and 94°6/ E, and flows for about 60 km through the floodplains before joining the Subansiri River ( a major tributary of the Brahmaputra) near Pokonighat at 27°1/ N and 94°6/ E (Dutt and Datta, 1976; Kaushik and Bordoloi, 2016). The Ranganadi catchment receives heavy rainfall during the pre-monsoon and monsoon period (Das and Ahmed, 2005).

Elevation within the basin varies greatly between the upper riparian hills and the lower riparian plains, ranging between 3746 m and 69 m. The slope varies from 0 to 83.02 degrees as can be seen in Fig. 2.1. The upper catchment of the river in Arunachal Pradesh is densely vegetated with mixed deciduous forest, while the lower catchment is comprised of sub-tropical forests (Das and Ahmed, 2005). The alluvial floodplains are quite fertile and largely cultivated with paddy as the primary crop.

Figure 2.2 displays the location of the Ranganadi dam and the examined extent of Ranganadi. Since the study concentrates only on the off-site impacts of the dam, particularly in the lower floodplain areas, the examined reach of the river begins approximately 25 km downstream of the RHEP dam at 27°20/ N and 93°59/ E, and ends at its confluence with the Subansiri. The total length of the examined reach is approximately 54 km.

26

Chapter 2

Fig. 2.1 Slope variation in Ranganadi.

Fig. 2.2 (A) The Ranganadi basin showing the river and the study length and location of the gauge station, (B) the diversion dam and reservoir as seen in Google Map and, (C) photo of the dam and reservoir collected during field visit.

2.3 Data and methodology

The hydrological alterations caused by the RHEP on downstream flows of Ranganadi were evaluated using the Indicators of Hydrological Alteration (IHA) model (Version 7.1) of The Nature Conservancy, USA (Richter et al., 1996), which assesses the degree of hydrological perturbations in a riverine system based on daily discharge data. The model

27

Chapter 2

uses 33 parameters, “organizedra into five groups, which provide information on ecologically significant features of surface and ground water regimes influencing aquatic, wetland and riparian ecosystems, to statistically characterize hydrological variation within each year.” Studies such as those by Magilligan and Nislow (2005) (21 gage stations across the United States), Graf (2006) (American Rivers), Fantin-Cruz et al. (2015) (Correntes River, Ponte de Pedra dam, Brazil), and Alrajoula et al. (2016) (Blue Nile River, Er Roseires Dam, Sudan) have successfully employed the IHA method to determine dam-induced changes in river flow. The IHA divides a single record of flow into pre-dam (also called pre-impact or pre-impoundment) and post-dam (post-impact or post-impoundment) periods based on the date of dam completion and returns ‘statistical summaries’ or ‘descriptive statistics’ of river flow before and after the dam (Richter et al., 1996; Graf, 2006).

24 years of daily discharge data pertaining to Ranganadi from 1991 to 2014 was obtained from the Water Resources Department of Assam with gauge/discharge (G/D) station at Pahumaraghat in North Lakhimpur District, located approximately 46 km downstream from the RHEP dam. This hydrological record was then used to assess the post-dam changes in downstream flow. The hydrological dataset was subjected to a Kolmogorov-Smirnov test for normality and the data was found to be a non-normal distribution at p < 0.005. Therefore, the hydrological alterations have been evaluated using non-parametric tests in IHA as well as in SPSS. Mean values have also been used at certain places, such as comparison between inter-annual flood flows or the inter-annual mean and peak flows, to correlate and describe the channel changes. The IHA parameters were separately compared for significant differences between the pre-impact and post- impact periods using rigorous statistical tests run in SPSS statistics (IBM SPSS Statistics 23). Mann-Whitney U test was employed to compare the different categories of median flow conditions obtained in IHA between the pre- and post-impact periods.

The morphological changes in the downstream river reaches have been assessed with the help of Geographical Information System (GIS) and Remote Sensing (RS). The use of GIS combined with aerial photographs and satellite imagery are a successful and essential tool for studying fluvial geomorphology (Ghosal et al., 2010; Aher et al., 2012; Sarkar et al., 2012; Kumar et al., 2015). “Complex patterns of bank migration are now commonly investigated based on temporal sequences of bankline data, captured from

28

Chapter 2

aerial photographs or remotely-sensed data” (Mount et al., 2013, pp 83). Aerial photographs were effectively used to investigate the impact of dams on river morphology by Gurnell (1997), Gilvear (2003) and Winterbottom (2000). Likewise, the data used for this study are primarily multi-temporal remotely sensed Landsat (MSS, TM, and ETM) satellite images for 15 individual years from 1973 to 2014 (obtained from the United States Geological Survey, i.e., USGS repository), spanning a time-frame of 41 years. Accordingly, the years (along with the month and date of the image) that were studied for channel changes are - 1973 (dated 11/15), 1976 (12/14), 1987 (12/03), 1989 (12/08), 1991 (11/12), 1995 (12/09), 2000 (12/06), 2002 (12/12), 2004 (12/17), 2006 (12/07), 2008 (11/26), 2009 (11/29), 2010 (12/02), 2013 (12/10), and 2014 (12/29). All the satellite images used in the study belong to the same season, i.e., the post-monsoon lean or dry period (principally November and December months) when the river experiences low- flow.

Channel boundaries and associated channel features were digitized at a scale ranging from 1:10,000 to 1:20,000 using the standard FCC band combinations. Digitization was done based on visual interpretation. The wetted perimeter, at the interface between land and water was delineated to construct the river bankline. However, barren or sparingly vegetated islands and sand bars were also included as part of the active river system. These morphological features are usually unstable in nature and are easily inundated by channel forming flows, i.e., bankfull discharges (Surian, 1999). Hence, two types of width were measured in this study – the bankfull width and the wetted width (Fig. 2.3A).

Bankfull channel width refers to the width of the river at the bankfull stage and includes width of the wetted channel as well as widths of the barren or sparingly vegetated islands and bars (mid-channel and point bars). Hence, the bankfull width of the river represents the total distance between the two banks of the river perpendicular to stream flow at bankfull discharge wherein the mid-channel and side-channel bars together with the flow transporting channels constituted the active area of the riverine system (Fig. 2.3A). However, side channels that bifurcate from the primary channel but do not rejoin the river were not included into the width calculations. Wetted channel width, on the other hand, is the sum of the widths of only the wetted channel (areas that show the presence of water even during the low-flow season), and excludes the bars and islands.

29

Chapter 2

For measuring channel width, the earliest year in this study, i.e., 1973 channel was divided into sections or transects at equal intervals of 1.5 km along the channel length perpendicular to the direction of river flow (Fig. 2.3B). Thus, widths were measured across 33 perpendicular sections (R1, R2, R3… R33) and average channel width calculated accordingly for all study years from 1973 to 2014. As different sections along the same river may exhibit varying types of morphological changes in response to hydrological or other factors, 5 larger reaches (or segments) were also created encompassing the various sections to measure the planform changes reach-wise (Fig. 2.3B).

Fig. 2.3 (A) Channel features that were distinguished following digitization of the river and how the bankfull and wetted channel widths of the river have been defined, (B) the 33 transects and 5 reaches that were distinguished to determine planform characteristics such as width, sinuosity and braiding.

The five reaches were created drawing from Graf’s (2006) description of ‘reaches’ as the length of a channel with ‘similar hydraulic, geomorphic, biological and chemical characteristics throughout the individual lengths’, such as an alluvial channel segment with similar meandering or braiding patterns. Therefore, the valley-wall constricted meandering part of the river from the starting point of the study length till section R4,

30

Chapter 2

where the river enters the floodplains, formed reach 1. Downstream from R4, where the river adopts a braided pattern till R10 constitutes reach 2. The middle-most floodplain part of the river with both sinuous and low braiding pattern was divided into two reaches – reach 3 (R 10 to R16) and reach 4 (R16 to R23). The lowermost meandering part of the river from R23 up till its confluence with Subansiri formed reach 5.

Like channel width, channel sinuosity, and braiding were examined individually for the entire study length as well as reach-wise, in order to capture both overall changes and the more specific spatial changes. Goswami et al. (1999) applied similar method of overall and segment specific measurement of channel sinuosity and braiding in their study of the Subansiri River. Channel sinuosity has been evaluated as,

SI = Lcmax/LR

where, LR is the length of the channel reach along a straight line, and Lcmax is the mid- channel length of the widest channel (Leopold and Wolman, 1957; Friend and Sinha, 1993, Dey, 2014). Channel braiding has been measured using Friend and Sinha’s (1993) braid-channel ratio definition, where braiding is a measure of channel multiplicity and -

B = Lctot/Lcmax

where, Lctot is the sum of the mid-channel lengths of all the channels in a reach and Lcmax as defined above.

For analyzing the bank migration processes of the channel, banklines of the individual years under comparison were superimposed on a GIS platform to generate migration polygons (Fig. 2.4). These polygons represent the actual areas that had undergone shifting or migration due to the lateral movement of the river’s banks between two study years. Area of each migration polygon was then calculated and the bankline shifting rate was determined using the following equation:

Rmb = (A/Lt1)/Y

where, Rmb is the migration rate; A is the area of the polygon; Lt1 is the length of the bankline at time t1 (previous year of the two years being compared) and Y is the number of years between the two years being studied. This method is similar to the system of calculating lateral movement of river channels over sequential years using the channel

31

Chapter 2

centerline as a unit of study. Giardino and Lee (2011) and Shields Jr. et al. (2000) had employed the said technique to calculate lateral migration of river channels with respect to the temporal and spatial movements of the channel centerlines.

Fig. 2.4 Right banklines of the Ranganadi channel in 2002 and 2014, where the grey polygons indicate the area of migration between the two examined years.

Thus, bankline migration was examined between the overall pre-dam and post- dam periods, i.e., between 1987-2002 and 2002-2014. Since 2002 was the year of dam completion and commissioning, it is used as a marker to divide the time-frames into pre- dam and post-dam periods. Similar temporal division was also applied to the analysis of river width. Bankline migration was also examined for periodic changes across the six time-frames – 1987-1991, 1991-1995, 1995-2002, 2002-2006, 2006-2010 and 2010-2014.

In addition to bankline shifting, the spatially distributed areas generated by the superimposed banklines of the river were further analyzed for changes in the erosion and deposition (or aggradation) patterns of the river and the associated impacts on the adjoining floodplain. Average annual rates of bank erosion and deposition at each bank were measured by dividing the area obtained at each polygon/segment by the number of years elapsed between the two years of study. Polygons where the lateral distance between the overlaid banklines was less than 30 m were ignored.

32

Chapter 2

2.4 Results and discussion

2.4.1 Alterations in river hydrology

The Brahmaputra Basin is characterized by a distinct monsoon season during which maximum amount of rainfall occurs in the region. Around 60% of the annual precipitation occurs during the period from June to September constituting the monsoon7 season (Ahmed et al., 2011). Notable amount of rainfall is also received during the pre-monsoon months of March to May (20-23% of annual rainfall). Rainfall decreases considerably during the post-monsoon season of October to November (6-8%) and is the least during the winter season (December to February, also referred to as the dry or lean season) with occurrence of only 2-3% of the annual rainfall (Ahmed et al., 2011). The calendar year beginning on January 1, of a year and ending on December 31, of the same year has been considered as the water year in this thesis to analyze the annual flow patterns. If the water year is constructed based on the distinct rainfall seasonality and not the normal calendar year, then the water year would start from March 1 of a particular year and end on February 28 or 29, of the next calendar year. Majority of the satellite images used in this thesis belong to the post-monsoon months of November and December. Therefore, the calendar year is being used in this thesis (to assess the changes in flow and their influence upon channel geometry) instead of the water year in order to reduce complexities arising from overlapping years.

Likewise, since the satellite images largely belong to the post-monsoon months of November or December, the lean period in the hydrological analyses has been defined by the post-monsoon months November to December (i.e., November 1 to December 31 of the same calendar year). This simplifies the comparison between low-flow conditions and channel parameters such as the wetted width determined from lean season images. The wet season, on the other hand, constitutes the period from June to September (June 1 to September 1 of the same calendar year).

Figure 2.5 presents the daily flow distribution recorded at the Pahumaraghat G/D station from 1991 to 2014, where the effect of the RHEP dam and its storage capacity can be witnessed post-2002. The hydrological data was subjected to non-parametric tests where median values were used to compare flow alterations. A Kolmogorov-Smirnov test

7 Also used synonymously as the wet season in this thesis. 33

Chapter 2

for normality revealed that there was a statistically significant difference between the two sets of hydrological data and a normal distribution (p = 0.000). Therefore, non-parametric tests were run to determine the pre- and post-impact flow changes. Nevertheless, mean flow conditions have also been presented for some pre- and post-regulation comparisons. The number of years (n) for both periods of study is 12.

Fig. 2.5 Pattern of daily flow in Ranganadi from 1991 to 2014 observed at the Pahumaraghat G/D station, Lakhimpur District, Assam.

(A)

(B)

Fig. 2.6 (A) Pre-dam (1987 to 2002) and post-dam (2003 to 2014) hydrographs of daily flow in Ranganadi (Pahumaraghat gauge station) displaying the difference between the two periods, (B) annual flow duration curves for pre-dam and post-periods obtained

from daily flow.

34

Chapter 2

The change in downstream flow following impoundment can be clearly observed in Fig. 2.6A, with distinct lowering of median daily flow. The annual flow duration curves between the pre-dam and post-dam periods also show that the difference is higher in the base flow conditions or flows with an exceedance probability greater than 70% (Fig. 2.6B). Likewise, the median annual flow decreased from 120 m3/s in the pre-impact period to 44 m3/s in the post-impact period, thus experiencing 63% reduction in downstream river discharge. The range of variation in median annual flow was also lower during the post-dam period, as it varied between 7 m3/s in 2013 and 94 m3/s in 2005; whereas it was comparatively higher during the pre-dam period, varying between 49 m3/s in 2002 and 187 m3/s in 1993 (Table 2.1).

Table 2.1: Annual flow statistics of the Ranganadi derived from daily discharge data from 1991 to 2014. Median Mean Maximum Minimum Std. Year discharge discharge discharge discharge deviation (m3/s) (m3/s) (m3/s) (m3/s) Pre-dam years 1991 144 198 151 846** 26 1992 72 106 79 479 31 1993 187 233 125 584* 38 1994 171 188 101 546* 59 1995 132 169 104 730* 49 1996 137 167 111 583* 57 1997 146 168 106 510 58 1998 180 215 131 626* 62 1999 105 125 75 388 23 2000 105 124 58 380 53 2001 69 78 23 171 54 2002 49 79 72 402 7 Post-dam years 2003 64 89 74 310 11 2004 91 99 57 277 29 2005 94 106 75 393 23 2006 62 78 57 259 20 2007 54 76 64 323 2 2008 21 42 53 695* 8 2009 68 104 81 416 23 2010 46 57 49 489 4 2011 11 43 69 480 1 2012 14 67 89 504 4 2013 7 21 37 189 2 2014 8 35 59 343 1 * = small flood peak, ** = large flood peak (as per IHA statistics)

35

Chapter 2

2.4.1.1 Group 1 alterations

As has been described by Richter et al. (1996), the 12 parameters in IHA Group 1 refer to the magnitude of monthly water conditions expressed in mean or median values and their corresponding coefficient of dispersion. The coefficient of dispersion in median indicates the ‘environmental contingency’, i.e. the extent of flow variation within a given month.

The IHA analyses revealed that flows had declined considerably across all 12 months in the post-dam period (Fig. 2.7A). The Mann-Whitney significance test indicated that median monthly flows between pre-dam and post-dam periods were significantly different and reduced in magnitude (Table 2.2). All 12 months exhibited higher than 50% decline in post-dam monthly median flows, with the change being highest in April. April flows had reduced significantly by 80%, from a pre-dam median of 123 m3/s to 25 m3/s post-dam (Mann-Whitney U = 17, p = 0.001). Significantly higher percentage of change was also observed in the post-dam monthly median flows of October at - 64% (U = 16, p = 0.001), November = - 64% (U = 12, p = 0.001), December = - 61% (U = 16, p = 0.001), January = - 63% (U = 6, p < 0.001), February = - 78% (U = 15, p = 0.001), March = - 74% (U = 16, p = 0.001), May = - 75% (U = 10, p < 0.001).

Alternately, the monsoon months of June, July, August and September displayed comparatively lower reductions at - 52% (U = 17, p = 0.001), - 52% (U = 19, p = 0.002), - 46% (U = 19, p = 0.002) and - 56% (U = 15, p = 0.001) respectively. Thus, the pre-monsoon and post-monsoon period exhibited higher alterations following dam obstruction. This decrease can be attributed to the phenomenon of water diversion to Dikrong even during the non-monsoon months (when the basin naturally experiences lower precipitation and thus lower flows), which has put more stress on water availability in the Ranganadi riparian zone.

With respect to the variation in flows within each month, coefficient of dispersion (CD) had increased markedly during the post-dam period for the months of January (+ 175%), February (highest at + 210%), April (+ 137%), May (+ 117%), November (+ 105%) and December (+ 152%). Once more, the monsoon months exhibited lower alteration, wherein July, September and October displayed an increase of 33%, 59% and 15% respectively. On the other hand, there was no change in post-dam CD in June and a decrease in August flows (-15%).

36

Chapter 2

Reduced dry season flow would lead to a reduction in the wetted area within the river channel and thus the wetted width during that period, a change that has been discussed in detail later in the chapter. Dam induced reduction in lean season flow would also result in elimination of water transport by the smaller secondary channels, thus reducing the active river width and subsequently impact upon habitat construction by certain riverine species.

Fig. 2.7 IHA parameters displaying the difference between the pre-dam and post-dam periods with respect to the - (A) median monthly (i.e., Group 1 parameters) and (B) annual extreme water conditions (i.e., Group 2 parameters).

2.4.1.2 Group 2 alterations

Group 2 constitutes 10 parameters that define the magnitude and duration of annual extreme (minimum and maximum) conditions of daily to seasonal durations. 7-day indicates the weekly, 30-day is monthly, and 90-day is the seasonal duration of flows. “For any given year, the 1-day maximum (or minimum) is represented by the highest (or lowest) single daily value occurring during the year; the multi-day maximum (or minimum) is represented by the highest (or lowest) multi-day average value occurring

37

Chapter 2

Table 2.2: Difference between the pre-dam and post-dam flow regimes of the Ranganadi following construction of the RHEP, obtained from IHA analysis of daily flow data from 1991 to 2014. The asterisks, ‘*’ and ‘**’, indicate significant differences between the two periods at 0.05 and 0.01 confidence levels (Mann Whitney significance test). Pre-dam Post-dam Alteration in median (%) Median Min Max Median Min Max Group 1 – magnitude of monthly flows January (m3/s)** 63 44 129 24 3 65 - 63 February (m3/s)** 60 17 124 13 3 98 - 78 March (m3/s)** 78 9 134 21 2 105 - 74 April (m3/s)** 123 37 235 25 2 96 - 80 May (m3/s)** 148 45 218 37 6 113 - 75 June (m3/s)** 198 93 388 94 22 158 - 52 July (m3/s)** 328 87 416 157 35 222 - 52 August (m3/s)** 230 82 380 124 47 274 - 46 September (m3/s)** 214 94 387 94 20 158 - 56 October (m3/s)** 144 73 248 52 9 127 - 64 November (m3/s)** 89 39 161 32 5 82 - 64 December (m3/s)** 70 25 129 27 2 77 - 61 Group 2 – magnitude of annual extremes 1-day min (m3/s)** 51 7 62 6 1 29 - 88 3-day min (m3/s)** 54 7 62 6 1 29 - 88 7-day min (m3/s)** 55 8 63 7 1 33 - 87 30-day min (m3/s)** 61 8 74 12 1 39 - 81 90-day min (m3/s)** 72 23 110 21 2 62 - 71 1-day max (m3/s)* 528 171 846 368 189 695 - 30 3-day max (m3/s)** 487 152 702 294 168 411 - 40 7-day max (m3/s)** 435 128 577 249 132 357 - 43 30-day max (m3/s)** 339 111 479 177 90 286 - 48 90-day max (m3/s)** 290 98 400 155 65 223 - 47 Base flow index 0.32 0.1 0.7 0.12 0.03 0.34 - 63 Group 3 – timing of annual extremes Date min (Julian date) 29 7 358 62 1 345 + 18 Date max (Julian date) 185 158 250 224 166 259 + 21 Group 4 – frequency and duration of pulses Low pulse count 7 2 15 8 4 17 + 14 Low pulse duration 4 2 17 4 3 21 0 (days) High pulse count 10 0 19 6 0 16 - 40 High pulse duration 4 2 6 2 1 4 - 50 (days)** Low Pulse Threshold = 68.63 m3/s, High Pulse Threshold = 204.7 m3/s Group 5 – rate (in m3/s) and frequency of daily flow changes Rise rate (m3/s/day)** 16 7 24 4 0.3 16 - 74 Fall rate (m3/s/day)* - 5 - 8 - 1 - 2 - 7 - 0.2 + 67 No. of reversals 129 70 150 109 74 154 - 16

38

Chapter 2

during the year” (Richter et al., 1996). This group also includes the number of zero-flow days and the base flow index. The base flow index is the ratio between the 7-day minimum flow and mean flow for the year (The Nature Conservancy, 2009).

In case of Ranganadi, analyses of the group 2 parameters revealed that the magnitude of both minimum and maximum flows have been significantly reduced post- impoundment (Fig. 2.7B). The decrease in daily to seasonal flows was significantly higher for all five categories of minimum flow, i.e., 1-day (– 88%, U = 9, p < 0.01), 3-day (– 88%, U = 7, p < 0.01), 7-day (– 87%, U = 7, p < 0.01), 30-day (– 81%, U = 7, p < 0.01), and 90-day (– 71%, U = 10, p < 0.01) minimum flows, compared to the decrease in maximum flows for the same duration (Table 2.2). At the same time, reductions in post- dam maximum median flows varied within a lower range of − 30% for the 1-day length to − 48% for the 30-day length. The median base flow index had also reduced by 63%, from 0.32 (pre-dam) to 0.12 in the post-dam (U = 19, p = 0.002).

This higher reduction in low flow post-impoundment, which was also visible in the annual FDC (Fig. 2.6B), is quite contrary to dam induced changes in low flow conditions in rivers elsewhere. Usually following impoundment, rivers tend to experience an elevation of low flows (daily to seasonal) (Magilligan and Nislow, 2005). However, in Ranganadi this may not hold true due to flow diversions even during the natural low flow season.

2.4.1.3 Groups 3, 4 and 5 alterations

Group 3 parameters denote the timing (Julian date) of the annual extreme flow conditions (annual 1-day maximum and minimum). Group 4 indicates the frequency and duration (in days) of high and low pulses in each year. Richter et al. (1996) defined hydrologic pulses as those annual occurrences when water conditions either exceed an upper threshold (i.e. above the 75th percentile of all daily values for the pre- impact period in case of a high pulse) or drops below a lower threshold (i.e. the 25th percentile for low pulse). The low and high pulse thresholds generated by IHA from daily flow values for Ranganadi from 1991 to 2002 are 69 m3/s and 205 m3/s respectively. These pulses essentially facilitate the exchange of nutrients, organic matter, and minerals between the river and its floodplain, while also influencing bed load transport (The Nature Conservancy 2009). Finally, group 5 denotes the rate and frequency of changes in the water conditions indicated as the

39

Chapter 2

number and median rate of positive and negative changes in consecutive daily flow conditions.

Of the nine parameters in the three above-mentioned groups, a statistically significant post-dam alteration was obtained in the duration of high pulses, the rise rate, and fall rate (Fig. 2.8). High pulses exhibited a decrease of 50% (U = 18.5, p = 0.005) in duration. Rise rate declined from a median of 16 m3/s/day to 4 m3/s/day (– 74%, U = 19, p = 0.002), while the fall rate increased from – 5 m3/s/day to – 2 m3/s/day (+ 67%, U = 33, p = 0.02).

Fig. 2.8 Temporal variation in the median high pulse duration (A), rise rate (B), and fall rate (C), of water conditions in Ranganadi between the pre-dam and post-dam periods.

2.4.1.4 Changes in high flows and floods

According to Wolman et al. (1957), a single high-flow can induce more changes in channel geometry and result in channel widening, than the low amount of change induced by many succeeding days of lower flow. Therefore, the channel forming flows or high flows were analyzed to examine their influence upon channel geomorphology. The IHA method distributes all flows that are greater than the base flow conditions in a river into three categories – high flow pulse, small flood, and large flood. High flow pulses indicate all channel flows that are typically bankfull discharges but do not overflow the banks. Small floods are flows that overtop the main channel and have a peak flow greater than 2- year return interval event. Large floods are the “more extreme and less frequent events”

40

Chapter 2

that have a peak flow greater than a 10-year return interval event (The Nature Conservancy, 2009). Both small and large floods usually influence the lateral movement of channels and formation of secondary channels.

Table 2.1 presents the annual mean and peak flows (annual maximum discharge) from 1991 to 2014. It can be observed that the mean flows display a decrease from 1999 onwards, which is prior to dam completion in 2002 (Fig. 2.9). Williams and Wolman (1984) did state that the usual passage of water and sediment in a river starts getting affected much before dam completion during the initial stages of construction itself. Water storage in the reservoir also starts before the river is dammed completely and electricity generation operations begin. This might explain the pattern of decrease in flow few years prior to complete impoundment. Additionally, the first unit (turbine) of RHEP was commissioned in March 2001, meaning that the diversion dam was completed well before that. So, flow obstruction in Ranganadi began prior to 2001 at the least, which likely explains the decline from 2001 onwards, following which flow continued to remain lower than the lowest pre-dam year.

Fig. 2.9 Pattern of decrease in median annual flow in Ranganadi from 1991 to 2014.

The annual peak flows can be divided into four distinct time-scales based on their changing temporal pattern (Fig. 2.10). From 1991 to 1998 the peaks range between 510 m3/s and 846 m3/s (barring 1992). The highest peak, at 846 m3/s, during the entire period of study from 1991 to 2014, is also recorded in 1991. The second time-scale is from 1999 to 2007, with relatively much lower peaks that range between 171 m3/s and 389 m3/s. The third time-scale is from 2008 to 2012 that shows a rise in the peak flows ranging between 416 m3/s and 695 m3/s. Peak flows reduce once again in 2013 and 2014 (fourth time- scale) reaching values comparable with the second time-scale. In general, the average

41

Chapter 2

yearly peak during the first time-scale was calculated as 613 m3/s, 323 m3/s for the second time-scale, 517 m3/s for the third time-scale, and 266 m3/s for the fourth time-scale. Therefore, a prominent attenuation of peak flow can be witnessed from 1999 to 2007, once again indicating a temporal change that starts a few years before and immediately following project commissioning.

Fig. 2.10 Variation in annual peak flows from 1991 to 2014, displaying the distinct lowering from 1999 to 2007 (Time-scale 2), i.e., a few years prior to and immediately following impoundment by RHEP.

The peaks of 1993 (584 m3/s), 1994 (546 m3/s), 1995 (730 m3/s), 1996 (582 m3/s), 1998 (626 m3/s) and 2008 (695 m3/s) were the small flood peaks with a 2-year return interval, while 1991 recorded the large flood peak with 10-year return interval time, as per the IHA statistics. Although, the 2008 flood that peaked on June at 695 m3/s was obtained as a small flood event in the IHA analysis, it caused large-scale devastation in the downstream riparian areas of Assam (Lakhimpur District) as it turned into a flash flood occurrence. According to reports procured from the North Lakhimpur Water Resources sub-division and the District Disaster Management Authority of Lakhimpur District, the June 2008 flash floods resulted in 9 embankment breaches along the river’s left bank, and 2 breaches along the right bank. The corresponding water level in the river at the time of the breaches was recorded as 96.86 m, which is higher than the recognized Danger Level (DL) of 95.02 m and High Flood Level (HFL) of 96.75 m for Ranganadi (at the Pahumara G/D station). The downstream flash floods had resulted from a combination of high precipitation during the monsoon season (and hence high inflow into the reservoir of the RHEP) and increased and sudden water releases from the dam.

42

Chapter 2

2.4.2 Changes in river morphology

2.4.2.1 The avulsion of 1991 and its repercussions on channel morphology

Goswami et al. (1999) described channel avulsion as the sudden shift in the position of a channel to ‘a new part of the floodplain (first-order avulsion)’ or the sudden capture of an old abandoned channel (second-order avulsion). The newly formed channel in due course of time may take the form of the original channel (Gogoi and Goswami, 2014), or may be abandoned once again with a shift back to the original channel. The latter was a significant change observed in Ranganadi, as discussed below.

One of the major changes to have taken place in the Ranganadi between 1973 and 2014 was the occurrence of channel avulsion and subsequent formation of an anabranch post-1991 (Fig. 2.11). Owing to repeated embankment breach in 1990 and 1991 between sections R26 and R27 on the river’s right bank; a few kilometers upstream of the river’s confluence with Subansiri, channel avulsion led to the formation of an anabranch to the west of the original Ranganadi channel (Fig. 2.11A and B). Following formation of the anabranch, river flow downstream from R26 gradually got diverted to the new channel, thus resulting in the complete abandonment of the old channel post-1995. The river took an entirely new course downstream from the embankment breach from 1996 to 2005, abandoning the old/original flow channel and displacing its confluence point with Subansiri River as well by approximately 5 km. However, the active water channel of the river does revert to its old channel of confluence with Subansiri in 2006 (Fig. 2.11E).

Owing to this avulsion and its ramifications on channel geometry for the period between 1995 and 2006, certain constraints were experienced with the pre-dam post-dam comparison of the planform features. The resultant modifications in methodology, analyses, and interpretation of the overall pre-dam post-dam and periodic variations in channel morphology are further explained in each of the planform characteristic sections.

2.4.2.2 Channel sinuosity and braiding

Channel pattern of a river can simply be defined as the plan view of the river as seen from above and can be broadly categorized as straight, sinuous, meandering or braided (Leopold and Wolman, 1957). Braided river channel pattern is characterized by the presence of multiple channels separated by frequently shifting bars and islands that are

43

Chapter 2

Fig. 2.11 (A) Breach in the right bank of the Ranganadi channel below section R26 in 1991 showing the flooded area, (B) subsequent avulsion and formation of a new channel in 1995, (C) the large migration polygon formed due to the shift in the banklines of 1991 and 1995 as a consequence of this avulsion, (D) 2002 image showing the abandonment of the old channel and diversion of flow into the new channel on the west, (E) activation of the original channel in 2006, and (F) reformation of the large migration polygon between the banklines of the 2002 and 2006 channel as a result of this return.

either un-vegetated or lightly vegetated (Leopold and Wolman, 1957; Lord et.al., 2009; Eaton et.al., 2010). However, more often rivers exhibit a gradual overlapping or merging between the different channel patterns and ‘do not easily fall into well-defined categories’ (Leopold and Wolman, 1957).

44

Chapter 2

The Ranganadi can be termed as a sinuous-braided river with an overall sinuosity index, SI, varying between 1.4 and 1.52 and braid-channel ratio, B > 1.2, based on the classifications by Leopold and Wolman (1957)8 and Friend and Sinha (1993). The river at reaches 2 and 3 (encompassing the sections from R4 to R16), is highly braided with multiple channels developed around both transient mid-channel bars and permanent island bars. B in reach 2 ranged between 1.8 in 2004 and 3.06 in 1991. Within reach 2, specifically across sections R5, R6, and R7, the planform displays two prominent side (or secondary) channels to the left of the primary channel separated by permanent island bars (Fig. 2.12A and B). Transient tertiary channels can also be seen during some years that

Fig. 2.12 The side channels between sections R5 and R7 that show a distinct water signature in the post-monsoon images of 2002 and 2010 (A and B), but are not clearly visible in the images of 2013 (C) and 2014 (D), suggesting lack of adequate flow in the river.

8 Leopold and Wolman (1957) classified rivers into three types of channel patterns based on the sinuosity index. A channel is straight when the SI is less than 1.05, sinuous when SI is between 1.05 and 1.5, and meandering when SI is higher than 1.5.

45

Chapter 2

fluctuates the braid-channel ratio across time. In case of reach 3, B ranged between 1.47 in 2010 and 2.03 in 2002, except in 1973, when B was lowest at 1.17. Table 2.3 shows the temporal and spatial changes in the sinuosity index and braid-channel ratio of the Ranganadi.

Unlike reaches 2 and 3, the channel at reaches 1, 4 and 5 exhibits a sinuous pattern with high sinuosity index and low braid-channel ratio (Fig. 2.13). Reach 1 is predominantly meandering with the highest sinuosity index (SI > 1.7) across all the examined years, mainly due to valley confinement located in the hilly terrain of the upstream riparian. SI in reach 4 varied between 1.39 in 2002 and 1.52 in 1987 and 1995, while in reach 5, SI varied between 1.35 in 1973, 1987 and 1995 and 1.65 in 2014.

Fig. 2.13 Lower reaches of the Ranganadi - reach 4 and 5, displaying a sinuous channel pattern.

From 1973 to 2014, the overall channel sinuosity varied between 1.4 in 1973 and 1.52 in 2004, while braiding varied between 1.23 in 2004 and 1.72 in 1995 (Table 2.3). There has not been much temporal change in overall channel sinuosity, but braiding exhibited a gradual decrease, especially during the post-dam years of 2004, 2011 and

46

Chapter 2

2014. Certain reaches exhibited more changes in channel pattern than the river as a whole. Reach 5 displayed an increase in sinuosity from 2002 onwards as can be observed from Table 2.3.

Table 2.3: Temporal and spatial variations in the sinuosity indices (SI) and braid-channel ratios (B) measured across the 5 reaches of the Ranganadi River as well as overall examined length from 1973 and 2014.

Years Overall Reach-specific

1 3 4 5 2 (starting (R10 to (R16 to (R23 till (R4 to R10) point to R4) R16) R23) confluence)

SI B SI B SI B SI B SI B SI B

1973 1.40 1.38 1.76 1.15 1.27 2.19 1.32 1.17 1.44 1.14 1.35 1.34

1987 1.43 1.41 1.72 1.06 1.33 2.11 1.14 1.70 1.52 1.34 1.35 1.11

1991 1.44 1.59 1.71 1.00 1.28 3.06 1.20 1.69 1.47 1.34 1.53 1.07

1995 1.46 1.72 1.69 1.00 1.33 2.11 1.14 1.70 1.52 1.34 1.35 1.11

2002 1.45 1.55 1.71 1.00 1.37 2.91 1.13 2.03 1.39 1.29 1.37 1.09

2004 1.52 1.23 1.78 1.05 1.40 1.80 1.16 1.65 1.49 1.02 1.44 1.00

2006 1.43 1.41 1.74 1.09 1.35 2.62 1.13 1.86 1.40 1.25 1.55 1.00

2010 1.43 1.25 1.71 1.00 1.32 2.07 1.12 1.47 1.45 1.33 1.57 1.00

2014 1.46 1.33 1.78 1.04 1.18 2.08 1.21 1.91 1.48 1.05 1.65 1.00

2.4.2.3 Disappearance of side channels across R5, R6, and R7 within reach 2

The river at reach 2 exhibits a highly fluctuating braid-channel ratio and a distinct decrease in sinuosity post-2010. This could mainly be due to the presence of the secondary channels that appear and disappear between different post-dam years, while the primary channel displayed a straightening pattern. Secondary channels are smaller active channels, also known as ‘side’ channels that traverse the floodplain along with the ‘mainstem’ channel, i.e., the largest active channel conveying the greatest discharge (Rapp and Abbe, 2003). Both vegetated and barren mid-channel bars and islands separate the mainstem channel and secondary channels from each other, thus forming a braided pattern. Braided channels are naturally rendered unstable due to the frequent temporal

47

Chapter 2

changes in the positions, size, and shape of the bars and islands. However, the pattern of frequent temporal changes in the braid-channel ratio of reach 2 in Ranganadi, also suggests an impact from hydrological alterations caused by the upstream dam.

From the satellite images, it can be observed that the historically existing side channels at sections R5, R6, and R7 of reach 2 become less prominent and cannot be distinguished post-2010 (Fig. 2.12C and D). The channels do not display a distinguishable water signature in the images of 2013 and 2014. This indicates either an absence of flow or extreme low flow that can only be attributed to the reductions in flow caused by the upstream dam, especially during the lean period of November to January. As a result, the island bar on the river’s left bank appears to have become continuous with the adjacent floodplain. Field visit to the site revealed that these side channels mostly remain inactive or partially active during the drier months due to highly reduced flow (that largely remain restricted to the primary channel), thus making the island bar continuous with the adjacent floodplain. Only during the monsoon period, when flow in the river increases, do these channels become active and convey water. However, the same channels are distinctly visible in the years prior to dam obstruction. The partial abandonment and irregular changes in the side channels also impact upon the bankfull width and migration pattern of the left channel bankline.

Rivers whose flows are reduced due to dam regulations typically display a simplification of the channel pattern. Multi-channeled braided alluvial rivers gradually transform into single-channeled sinuous rivers. The Ranganadi has also displayed a slow yet visible transformation from braided to a more sinuous channel, principally at the lower downstream reaches.

2.4.2.4 Changes in channel width

Channel width of a river displays continuous adjustment and variation to the water and sediment flow changes in the river (Church, 1995; Surian, 1999). Width of a river undergoes both temporal and spatial variations depending upon the flow and sediment conditions. Here I present width measurements across different reaches of the Ranganadi as well as the average width of the river calculated from the sectional widths.

48

Chapter 2

Average channel widths (bankfull and wetted) were, however, affected by the 1991 avulsion and the ensuing change of river course downstream of section R26. Thus, the pre-dam and post-dam width comparison was done based on two different types of channel from section R27 until the confluence with Subansiri. From 1973 to 1991, the sectional widths at R27, R28, R29, R30, R31, R32 and R33 were calculated for the original channel. However, from 1995 to 2005, channel width at these sections was calculated for the avulsion channel since original channel was abandoned during those years. From 2006 to 2014, when the river shifted back to the original channel, widths at the given sections were again calculated for the original channel. Thus, width of the river from 1995 to 2005 is for the avulsed channel in the lowermost reach of the river. The comparison of width changes between pre- and post-dam periods has been done considering this variation.

Figure 2.14 demonstrates the temporal variation in bankfull and wetted channel width of the Ranganadi from 1973 to 2014. Average bankfull width of the river varied between 214 m (± 255, n = 33) in 2013 and 423 m (± 326, n =33) in 1973 (Table 2.4). Comparison between the two-periods revealed that the post-dam bankfull channel had reduced by 24%. Bankfull width in the pre-dam period (1973-2002) was calculated at an average of 356 m (± 39, n = 8), but reduced to 277 m (± 44, n = 7) in the post-dam period (2003-2014), indicating a narrowing pattern that was more pronounced during the post- dam period. Furthermore, within the post-dam period higher reductions in width were witnessed during the later years of 2010, 2013, and 2014. From 1989 to 2009, width remained stable fluctuating within 300 m (2008) to 348 m (2002), but decreased below 300 m from 2010 onwards. Average width also became quite low in 2006 at 256 m.

In terms of channel width enclosed by the wetted area of the Ranganadi (i.e., the wetted channel width), it ranged between an average of 242 m (± 145, n = 33) in 1976 and 79 m (± 36, n = 33) in 2014. Significant difference was observed in the average wetted widths between the pre-dam and post-dam periods with 38% reduction in post- dam width. Average wetted width for the pre-dam period was calculated at 183 m (± 27, n = 8), while the same for the post-dam period was found to be 113 m (± 27, n = 7). However, unlike the bankfull width, much fluctuation could be observed in the pattern of decrease in the wetted width amongst the post-dam years. The years immediately following dam completion saw greater fluctuation and reduction in the wetted width

49

Chapter 2

measured at 109 m (± 44, n = 33) in 2004 and 96 m (± 73) in 2008. Alternately, 2006, 2009 and 2010 recorded comparatively wider wetted channels averaging at 137 m (± 88), 142 m (± 123) and 141 m (± 79) respectively; though still narrower than all pre-dam years (Table 2.4). Further reduction was witnessed in 2013 and 2014 at 86 m (± 34) and 79 m (± 36), which was witnessed in case of the bankfull width as well. Downstream flows during 2013 and 2014 had also declined sharply (Table 2.1), and it is highly likely that the two types of reduction are related to each other.

(A)

Bankfull

Width (m)

(B) Wetted

Width (m)

Fig. 2.14 Temporal variation in the bankfull (A) and wetted (B) channel widths of the Ranganadi from 1973 to 2014.

Different reaches along the same river often exhibit varying widths. Similarly, in Ranganadi, the river is widest at its braided-reach, i.e., reach 2 (> 500 m), followed by reach 3 (> 240 m) and reach 4 (> 192 m). The meandering part of the river, i.e., reaches 1 and 5 are comparatively much narrower. A reach-wise analysis of width variation reveals that bankfull width across reaches 3, 4 and 5 displayed notable post-dam decrease particularly post-2004 (Fig. 2.15).

50

Chapter 2

Table 2.4: Average, maximum and minimum bankfull and wetted widths of the Ranganadi from 1973 to 2014 obtained from 33 sections along the study length of the river.

Year Bankfull width Wetted width Average Maximum Minimum Average Maximum Minimum Stdev Stdev (m) (m) (m) (m) (m) (m) 1973 423 326 1295 77 200 96 412 45 1976 369 287 1182 90 242 145 712 46 1987 401 272 1087 65 158 92 496 57 1989 314 238 1141 46 170 105 464 44 1991 339 243 1218 62 183 113 458 53 1995 326 289 1204 65 180 115 454 31 2000 325 288 1215 46 173 102 424 40 2002 348 290 1225 70 161 85 439 44 2004 339 266 1244 57 109 44 201 37 2006 256 283 1246 31 137 88 348 31 2008 300 324 1379 43 96 73 264 0 2009 315 359 1567 35 142 123 518 35 2010 273 285 1092 49 141 79 378 49 2013 214 255 1127 43 86 34 207 41 2014 238 271 1156 39 79 36 226 25

Fig. 2.15 Temporal variation in the average bankfull width across the five reaches in Ranganadi from 1973 to 2014.

Modification of channel width in Ranganadi was observed to be quite complex and marked with fluctuations affected by formations and alternately abandonment of side channels between different years, incidences of avulsions and overbank sand casting. Higher reduction in both bankfull and wetted widths during 2013 and 2014 was strongly influenced by the absence of detectable side channels across sections R5, R6 and R7

51

Chapter 2

(likely due to absence of flow or presence in minimal amount). Changes in point bar formation also affected bankfull width, as they increased the width during some years while reducing it during others. Wetted width of a river varies greatly according to the stream flow conditions during a particular day. All the images used to measure channel width in this study belong to the post-monsoon months of November and December (dated between 12th November and 29th December), when flow conditions in the river remain fairly similar (exhibiting low flows) with no large variations and hence comparable between the years. Therefore, a decreasing trend in wetted width during the post-dam period can be attributed to overall flow reductions during the same period by the upstream dam. Particularly, in 2008, it was observed that the river from R14 to R29, exhibited a discontinuous water channel (Fig. 2.16).

2.4.2.5 Absence of a visible water signature in the downstream reaches of the 2008 channel and its impact on wetted channel width

Water covered areas in a satellite image can be distinguished from the blue colour or signature that appears in a False Colour Composite (FCC) band combination. Vegetation appears as red (with different shades of red depending upon its density) and sand bars as bright white. The absence of a visible water signature, therefore, likely indicates the absence of flow or extremely low flow in the channel. Since the wetted width is calculated by taking into consideration only the water covered portions of the river channel, sectional width becomes 0 if there is an absence of water signature at the particular section of the river.

The Ranganadi primarily exhibits a single-channel meandering planform downstream from section R14. In the 2008 satellite image (26th November), the river lacks a distinctly visible continuous water channel downstream from section R13, and water signatures can be seen only in patches (Fig. 2.16). The channel appears completely dry across sections R20, R21 and R26, thereby recording 0 widths at the said sections. Hence, the highly reduced average wetted width of 2008 at 96 m. On the other hand, the bankfull width during the same year remained unaffected since the outline of the channel could be detected from the bright white signature of the dry riverbed. Near section R13, the channel also displays a breach on the left bank, which could have led to lesser flows passing to the lower sections. An absolute drying of the water channel was observed only

52

Chapter 2 in the 2008 image and hence, can be considered as a one-time phenomenon, unrelated to dam-induced reductions in downstream flows.

Fig. 2.16 The fragmented water channel from section R17 to downstream that shows the river to be dry and devoid of flow in 2008 (post- monsoon image dated 26th November).

2.4.2.6 Channel width calculated up till section R26

To eliminate any anomalies or inadequate comparisons of river width arising from the presence of two different primary channels below section R26, average width (bankfull and wetted width) was separately calculated for a shorter stretch of the river, i.e., only till R26 (from the starting point) involving a single un-abandoned primary channel.

Consequently, the river still exhibited a narrowing pattern in accordance with the pattern of temporal change for the complete length of the study reach. The percentage of decrease in post-dam average width of the channel till section R26 was calculated to be 19% in case of the bankfull width and 38% for the wetted width. Decrease was again found to be higher for the wetted channel than the bankfull channel.

These significant reductions in the channel width of the Ranganadi (especially for the wetted channel) clearly indicate the temporal and spatial adjustments undergone by the river’s planform as a response to dam-induced alterations of its natural flow regime.

53

Chapter 2

Width of a river undergoes active changes with changing flow regimes along with the added impetus of human interventions like dam structures, which is clearly evident in the case of the Ranganadi.

2.4.3 Bankline migration, erosion, and deposition patterns

Bankline migration in an alluvial river may simply be defined as the change in the position of the banklines across time and space. The direction and amount of shift usually vary along different sections of the river and between the two banks, a phenomenon that has been reported in the Subansiri River9 by Sarma and Phukan (2006). In the braided multi-channel sections of a river, both banks may simultaneously shift towards opposite directions, thereby causing channel widening10. These shifts occur through various channel changes. Erosion and avulsion are the two most common processes resulting in bankline migration. As the Ranganadi flows from the north to south, the banks move either to the west or to the east. Hence, bankline migration, erosion, and deposition have been measured for the two banks separately. Similar method was used by Sarma (2006) to study bank erosion and bank shift by the two respective banks in the Brahmaputra. In Ranganadi, the lateral movement of the right bank to the west indicates erosion and movement to the east indicates deposition. Similarly, the left bank moving to the east denotes erosion of the adjacent floodplain while movement to the west denotes deposition. Migration has been studied in terms of both periodic changes as well as the overall shift between 1987 and 2002, which corresponds to 15 years of unaltered pre-dam period; and between 2002 and 2014 corresponding to 12 years of altered post-dam period.

The 1991 avulsion phenomenon, once again affected the measurement and temporal comparison of bankline migration and the associated processes. While 1987 and 2014 banklines were defined by the original channel of Ranganadi, 2002 bankline was demarcated by the avulsion channel from section R26 to the confluence with Subansiri (Fig. 2.11D). As analyzing the pattern of bankline shifting involves superimposing the channel banklines of the two studied years, abnormally large migration polygons with areas ~ 25 km2 were thus obtained from overlaying the banklines of 1987 with 2002 and

9 The Subansiri is one of the largest tributaries of the Brahmaputra and neighbors Ranganadi and Dikrong. The three display similar alluvial features and channel processes although on different spatial scales. 10 This process was witnessed in the braided belt of the Dikrong River, which is discussed in the next chapter.

54

Chapter 2

2002 with 2014 (Fig. 2.17). Even though this area denotes the change in the positions of the active channel of the river from 1991 to 2006, its formation however did not take place through a gradual westward lateral shift of the banklines. Therefore, the area of this polygon was not incorporated in calculating channel erosion and deposition, since it does not signify any of those processes. However, concerning bankline migration (which is denoted by the change in the position of the banklines owing to lateral movement as well as avulsions), the area, and the rate of shift obtained for the given migration polygon (downstream of R26) was included. The sectional shift rate obtained at the said avulsion polygon did raise the average migration rates to very high values in case of both overall and periodic bankline migration calculations.

Fig. 2.17 The large migration polygon that was generated during the examined periods of 1987-2002 (pre-dam) and 2002-2014 (post-dam) due to the 1991 avulsion and subsequent shift of primary channel below section R26. Both banks of the river, thereby, shifted to the west and remained so, till their return back to the original path of flow in 2006. The avulsion polygon between 1987 and 2002 indicates the westward shift of banklines, while the one between 2002 and 2014 indicates an eastward shift of banklines due to the aforementioned return.

55

Chapter 2

2.4.3.1 Overall bank migration between 1987-2002 and 2002-2014

Average migration rates (Rmb) by the two banks did not vary much between the two

periods, as can be seen in Table 2.5. From 1987 to 2002, the average Rmb by the right bankline was 9.9 m/y and 8.7 m/y by the left bankline, which changed to 9.3 m/y and 10.2 m/y respectively from 2002 to 2014. The standard deviations of these average rates were exceptionally high owing to the inclusion of the large sectional rate (also the maximum) at the avulsion polygon below R2611 (Table 2.5). The high standard deviations could also have resulted from the wide spread between the sectional migration rates during both periods of study, ranging from ~ 1 m/y to more than 125 m/y. Consequently, the right bank displayed a maximum migration rate of 123.5 m/y (pre-dam) and 138.3 m/y (post- dam). The left bank showed a maximum migration rate of 131.5 m/y in the pre-dam and 151.4 m/y in the post-dam. Both banks displayed higher maximum migrations rates over the post-dam period. It was observed that the rates of migration between the two banks differ and was not uniformly spread. The amount and rate of erosion and deposition by the two banks during both periods were also unequal (Table 2.5).

Changes in the position of the river banklines mostly occurred through a combination of -

(i) Erosion induced lateral migration of the banks,

(ii) Avulsions resulting in formation of new side channels (or anabranches) and consequent reconstruction of the river banks and,

(iii) Avulsion induced shift in the position of the main channel and development of an entirely new primary channel (that changes the river course completely as witnessed post- 1991 below R26),

(iv) Abandonment of side channels, and

11 The average migration rates calculated by excluding the sectional rate at the mentioned avulsion polygon, stood at 5.8 m/y (± 6.5, right bank) and 5.6 m/y (± 6, left bank) in the pre-dam period and at 5.8 m/y (± 7, right bank) and 6.1 m/y (± 7.2, left bank) in the post-dam period. The corresponding maximum migration rates (recorded at different spatial locations than the avulsion polygon) were observed to be 30 m/y and 22 m/y respectively for the right and left banklines during 1987-2002; and 33 m/y (right bank) and 39 m/y (left bank) during 2002-2014. The sectional rates during both time-periods still varied within a large range of ~ 1 m/y to more than 20 m/y for both banks.

56

Chapter 2

(v) Development of mid-channel and point bars along the channels.

Table 2.5: Overall bankline migration, erosion, and aggradation between pre-dam (1987- 2002) and post-dam (2002-2014) time-periods.

1987 – 2002 (15 years) 2002 – 2014 (12 years)

Right bank Left bank Right bank Left bank Average rate of bankline 9.9 8.7 9.3 10.2 migration (Rmb) (m/y) Stdev 22.4 20.6 22.6 25.2

Maximum Rmb (m/y) 123.5 131.5 138.3 151.4

Minimum Rmb (m/y) 1.3 0.7 1.1 1.1

Total area of erosion (km2) 3.2 2.1 1.9 0.7 Annual rate of erosion 0.2 0.1 0.2 0.1 (km2/y) Total Area of aggradation 2 2.7 2.6 4.2 (km2) Annual rate of aggradation 0.1 0.2 0.2 0.3 (km2/y)

Rmb = rate of bankline migration, Stdev = standard deviation

Besides the avulsion polygon below R26, left migration during the pre-dam period was notably high across the polygon at R11, where the bankline had shifted westward 2 over an area (A) of 0.8 km (Rmb = 22.4 m/y) causing channel straightening (Fig. 2.18A).

The left bank also exhibited prominent shifting across sections R19 (Rmb = 21 m/y, A = 2 2 0.5 km ) and R20 (Rmb = 20 m/y, A = 0.7 km ), where there was a mirroring of the meander bends (Fig. 2.18B). The former resulted in aggradation along the left floodplain due to a westward shift of the left bankline, while the latter resulted in erosion due to an eastward shift. Erosion, here, has been considered as an expansion of the active belt of the river by lateral migration.

The post-dam period recorded high migration by the left bank at sections R5, R6, and R7 – the braided belt of the river, characterized by the two prominent secondary channels 2 (Fig. 2.18C). The large westward shift in the position of the bank over 2 km (Rmb = 38.8 m/y), however, did not take place through a gradual lateral movement of the primary

57

Chapter 2

channel, but due to the partial abandonment of the secondary channels. The same phenomenon was also responsible for the distinct lowering of the braid-channel ratio and width at the particular reach12. Flow reduction by the upstream dam is evidently the main cause of this channel abandonment. If we compare the timing of channel abandonment and flow alteration, it is quite clear that the mean annual and dry season discharge decreased considerably post-2011 (Table 2.1). Whatever volume of water flows down to the river at reach 2 is transported through the primary channel alone, leaving the side channels dry or with negligible flow. The post-dam period also saw increased aggradation by the left bank with a total area of 4.2 km2 as opposed to 2.7 km2 pre-dam. Erosion, on the other hand, reduced from 2.1 km2 to 1.7 km2.

If the capture of the Joyhing channel is taken into consideration than left bank migration post-dam would be quite large across the lower sections from R8 to R13. However, the Joyhing is not an anabranch formed by Ranganadi on its own; rather a tributary captured by the latter through avulsion, and therefore does not truly signify shifting of the bank for the main river under study.

With respect to migration by the right bankline, the highest rate during the pre-dam period was once again obtained at the avulsion polygon below R26. The second highest rate of migration was measured at 30.2 m/y, occurring at the same polygon that was formed due to the change in the meander bend across R20 marked during the pre-dam left bank migration (Fig. 2.18B). The eastward shift over an area of 0.81 km2 had resulted in deposition along the right floodplain and erosion along the left plain.

13 2 During the post-dam period, maximum Rmb at 33.4 m/y (A = 1 km ) occurred across R14, due to the abandonment of a temporary side channel formed through avulsion. Another notable shift by the right bank during the post-dam period was the avulsion of the primary channel to the west across sections R7 and R8 (Rmb = 26 m/y, A = 1.2 km2) resulting in erosion of the adjacent Dejoo tea estate. Total area of erosion by the right bank was lower in the post-dam period at 1.6 km2 against 3.2 km2 pre-dam, while deposition did not change much.

12 The phenomenon of secondary channel abandonment at reach 2 (sections R5, R6 and R7) has been elaborated earlier in section - ‘2.4.2.3 Disappearance of side channels across R5, R6 and R7 within reach 2’. 13 Besides the avulsion migration below R26.

58

Chapter 2

Fig. 2.18 Bankline migration patterns in Ranganadi during the pre-dam and post-dam periods, (A) Channel straightening across R11 that resulted in the largest migration by the left bank (in a westward direction) between 1987 and 2002, (B) mirroring of meander bends across R19 and R20 resulting in aggradation along the left floodplain due to westward shift in the former and aggradation in the latter due to eastward shift of the left bankline, (C) prominent shifting of the left bankline to the west from R5 to R7 during the post-dam period (i.e., 2002-2014) due to partial abandonment of the side channels.

59

Chapter 2

Therefore, it can be concluded that even though the mean migration rate of the river remained relatively unchanged across time, specific reaches had undergone large migrations while other reaches remained, more or less, stable. Such reaches may exhibit conditions that are more vulnerable to flow alterations.

Bank migration is a complex process caused by a combination of multiple factors other than discharge alone. Slope, sediment load, and nature of the bank materials are also important controls of bank migration and erosion. However, the influence of these variables has not been studied in this thesis due to lack of data and other logistic constraints. It is also difficult to separate the anthropogenic causes such as dams from the natural causes of bank shifting given that rivers such as Ranganadi and Dikrong are naturally quite dynamic. High monsoon discharges accompanied by large sediment loads increase the instability of these rivers and cause irregular shifts in the position of the primary as well as side channels within the active belt of the river.

2.4.3.2 Periodic bankline migration

In order to obtain a more in-depth picture of the bank migration, temporal changes in the position of the river’s bankline were studied separately for the six shorter periods of 1987-1991, 1991-1995, 1995-2002, 2002-2006, 2006-2010 and 2010-2014. Similar to the anomalous migration polygon obtained during 1987-2002 and 2002-2014 due to the 1991 avulsion and subsequent formation of a new primary channel in the following years till 2006, superimposed banklines of 1991-1995 and 2002-2006 periods also brought up anomalies. Banklines in 1991 and 2006 were outlined by the original channel of Ranganadi, while 1995 and 2002 bankline was formed by the avulsion channel. Hence, the superimposed banklines of 1991-1995 and 2002-2006 gave rise to the large migration polygon downstream of R26 (Fig. 2.11C and F). The exceptionally high sectional shift rate at the given polygon skewed the average periodic shift rate between the above- mentioned periods towards anomalously high values.

Table 2.6 displays the various migration rates, area of erosion and deposition during different time-periods, before and after dam construction. Between the two banks, migration by the left bank was slightly higher than the right bank across all time-periods. The post-dam periods of 2006-2010 and 2010-2014 displayed comparatively higher average migration rates by both banks than the pre-dam periods. Migration in 2010-2014

60

Chapter 2

at 12 m/y (right bank) and 15 m/y (left bank) was greatly influenced by the abandonment of side channels between R5 to R7. The sections or polygons that underwent maximum bank migration during each of the six time-periods are listed in Table 2.6, and show that the length of the river from section R5 to R20 encompassing reaches 2, 3 and 4 was most affected (other than the avulsion polygon). The migration rates varied notably from one period to another. Erosion and deposition also varied greatly between different time- periods.

Analysis of the overall and periodic bank migration patterns of the Ranganadi suggests that so far the flow alterations caused by the upstream dam has neither drastically increased nor decreased the rate of migration. Usually, under a reduced flow regime following impoundments, rivers exhibit decreased channel or bank migration as has been documented in the Missouri River (Montana, USA) and Willamette River (Oregon, USA) (Shields Jr. et al., 2000; Wallick et al., 2007). Different bank protection and erosion control measures such as embankments, spurs, dykes, levees and other revetments also constrict the lateral movement of the river to a large extent. Such features similarly restrict channel widening as they reduce the space for the river to expand.

Gendaszek et al. (2012) reported narrowing of the river channel by nearly 50% and transformation from a previously wide, anastomosing channel to a simplified single- channel planform in case of the Cedar River (USA) owing to flow regulation by dams and bank stabilization structures such as revetments. Channel migration rates were also found to be lower in the confined reaches of the regulated river as compared to the unconfined reaches. Post-regulation high flows, despite being lesser in magnitude than the pre- regulation high flows, effectively formed geomorphic features such as side channels, gravel bars and pools in the unconfined regulated reaches, thus indicating the impact of revetments. The influence of bank protection measures in resisting channel migration and erosion have also been discussed by Curran and McTeague (2011) with respect to the Matanuska River, Southcentral Alaska.

A similar phenomenon can be suggested in case of Ranganadi where embankments and other bank stabilization structures could have worked together with flow regulation to induce channel narrowing. However, the high maximum migration rates across specific sections suggest the failure and breaches of such structures along

61

Chapter 2

ted 3

15 16 87 R7 0.3 LB 0.75 0.53 0.13 R5 toR5 2014 2014

(Y=4) 3 2010 - 12 11 62 RB R10 1.26 0.31 0.53 0.13

3

11 10 50 0.8 0.2 1.3 LB R19 0.33 2010 2010

(Y=4) 3 11 16 2006 - 1.8 1.7 RB 105 R14 0.45 0.42

9 7 3

35 0.3 LB R10 0.57 0.14 1.19 R9 toR9 2006 2006

(Y=4) 8 5 4 to 24 2002 - RB R15 R15 R18 0.07 0.02 1.27 0.34

2 9 9

45 LB R11 0.36 1.88 0.27 2.53 2002

(Y=7) 7 9 2 55 1995 - RB R20 2.57 0.37 1.65 0.24

9 8 3

43 R8 LB 0.98 0.24 1.29 0.32 1995 1995

(Y=4) 8 6 3 31 1991 - RB R11 1.34 0.33 0.57 0.14

9 2 11 64 LB 1.99 0.52 0.36 0.16 and R9 across R8 R8 across 1991 (Y=4)1991

3 1987 - 10 10 72 1.1 RB 0.73 0.18 0.28 and R8 between R7 R7 between

) 2 /y)

2 ) 2 /y)

2

(m/y) (m/y) (m/y) (m/y) mb mb mb mb

deposited area area (km deposited Periodic bankline migration, erosion and deposition patterns of the Ranganadi between 1987 and 2014 along with the most affec the with along 2014 and 1987 between Ranganadi the of patterns deposition and erosion migration, bankline Periodic Table 2.6: channel sections. Stdev R Maximum Section(s) exhibiting Section(s) exhibiting R maximum Rate of erosion (km erosion Rate of R Average bank left = LB bank, right = RB deviation, standard = Stdev migration, of rate = Rmb Rate of deposition (km deposition Rate of Total Minimum R Minimum Total eroded area (km area eroded Total

62

Chapter 2

certain structures, which was evident during the 2008 flood. Embankments, despite being a common erosion protection feature along rivers in Assam, are only partially successful in controlling erosion and overbank flooding. Breaches may occur due to weakening of the structure and poor maintenance coupled with high flood discharges. The resulting flash flood in the adjacent floodplain and sand casting can cause more damage than a normal overbank flood. The breaches during 2008 and 2009 on the left bank of Ranganadi in its braided belt at reach 2, had caused not only major planform changes (such as the capture of the Joyhing channel), but also long-term socio-economic impacts on the agrarian riparian communities residing in the affected area.

However, once breaches are sealed and the embankment repaired, the river is restored back to its original path of flow and any previously formed channels are abandoned. Vegetation usually takes over such channels in the absence of water and especially if the inflow into the avulsed channel was solely from the main river. For instance, the formation of the avulsion channel downstream of R26 at reach 5 (right bank) in 1995 that displaces the primary channel for 13 km is primarily a result of embankment failure in 1990 and 1991. Once the embankment break was sealed, flow shifted back to the original channel.

2.4.4 Capture of the Joyhing channel and diversion of flow in the post-dam period

A significant channel change in the post-dam period was the re-capture of the old channel of the Joyhing River. Joyhing is one of the tributaries of the Ranganadi that originates in Arunachal Pradesh and joins the river on its left bank, in the floodplains of Assam. The confluence point between the two channels has shifted multiple times in the past. The old Joyhing channel ran parallel to the Ranganadi on its left bank for approximately 9 km before joining the latter between sections R13 and R14 (Fig. 2.19). However, in 1987 the confluence shifts upstream by approximately 9 km (near section R7) as a result of both head-ward erosion by the Joyhing channel (aided by a steep gradient) and past flooding and sand casting by the Ranganadi upon Joyhing. The new point of confluence closed the previously existing narrow gap of ~ 600 m between the two rivers. On the other hand, the original Joyhing channel still persisted as a secondary branch and continued to flow downstream. So, the Joyhing joined Ranganadi at two points along the length of the main river. With time, as most of the flow from Joyhing joined Ranganadi at its upstream confluence near R7, the rest of the old channel grew smaller until it turned into a narrow

63

Chapter 2 drain with low depth and low flow. Hence, it is not prominently visible in the satellite images from 1988 to 2011(Fig. 2.19). Riparian life along the Joyhing also adapted to a shrunken low flow channel that did not cause much trouble, but provided an important source of drinking and domestic water and fish protein.

b = bifurcation in the Joyhing channel that flows down to join the Ranganadi at C1,

C1 = first confluence point between Ranganadi and Joyhing,

C2 = second confluence point,

* = No. 1 Pachnoi Ujani village

Fig. 2.19 Temporal and spatial transformation of the Joyhing channel that joins the Ranganadi on its left bank at two points (C1 and C2). The satellite images shown in the figure belong to 1987, 2000, 2010 and 2014 (scale = 1:40000) and it can be observed that the Joyhing is much smaller in size and not clearly visible in the pre-dam images (1987 and 2000), whereas the same can be clearly distinguished in the post-dam images of 2010 and 2014.

64

Chapter 2

However, the channel once again underwent modification in 2008 following occurrences of flash floods and course alteration. Channel avulsion had resulted in the diversion of flood flow into the old Joyhing channel, but it was only by 2010 when the smaller channel was completely recaptured and transformed into an anabranch. In the images of later years, the anabranch becomes distinctly visible and appears to have grown in size (being reworked by the water and sediment flow of the main Ranganadi River), transporting substantial amount of flow and sediment from the main river (Fig. 2.19). The term, ‘recaptured’ is used since flows from the main river had spilled into Joyhing in the past, i.e., in 1973 as well. Hence, the old Joyhing channel and the area enclosed in between the two rivers fall within the historically active zone of influence of the Ranganadi. Additionally, the left bank embankment of the Ranganadi actually runs along the left bank of Joyhing, meaning that the smaller channel is enclosed within the space allotted for the main river to expand. The fluvial dynamics of the current Joyhing anabranch and its increasing size post-2008, however, remains greatly influenced by the upstream flow regulations of the Ranganadi dam.

Fig. 2.20 The dry channel bed of the Joyhing (A) and the bamboo bridge (B) that connects the mid-channel island bar, caught between the channel and Ranganadi, with the left bank

floodplain of the latter and the district headquarters. The images shown in the figure were collected during field visit in December 2013 and the bridge is located near No.1 Pachnoi Ujani village on the island.

Field surveys of the site in December, 2013 had revealed that there exists a sharp contrast in seasonal flow in the channel. The channel remains completely dry and devoid of any flow during the lean season and alternately swells up with heavy flow during the wet season months. The formation of the anabranch resulted in the establishment of an island bar with an area of more than 5.2 sq. km. This distinct vegetated island supports

65

Chapter 2

huge tracts of fertile agricultural lands and numerous wetland ecosystems within the principally agricultural villages of Pachnoi Ujani No.1 and No.2 and the Kalabil block. One of the impacts upon the residents of these villages is that during the wet season, more than the main Ranganadi, the Joyhing tributary floods the area and cuts off communication with the mainland with no permanent concrete bridge to cross over (Fig. 2.20). Chapter 4 discusses in detail the various socio-economic impacts faced by the downstream riparian villages in the aftermath of the 2008 floods and other post-dam changes in the river hydrology and morphology.

2.4.5 Comparison between flow conditions and channel width

In order to find out the influence of flow upon channel width, Spearman’s rank correlation coefficients (rs) were evaluated between different conditions of flow, i.e., yearly mean, maximum and minimum flow, the April-May mean flow (pre-monsoon flow), the June-September mean flow (monsoon flow) and channel width (average, maximum and minimum bankfull and wetted widths).

The average bankfull channel width exhibited significant positive correlation with

the yearly mean (rs = 0.76, p = 0.007) and minimum flow (rs = 0.62, p = 0.04). It also

exhibited significant correlation with the pre-monsoon and monsoon mean flows at rs =

0.72 (p = 0.01) and rs = 0.73 (p = 0.01) respectively, but did not display a significant

correlation with the annual maximum flow (rs = 0.47, p = 0.14). No significant relation was found between the maximum bankfull width and any of the flow types.

The average wetted width demonstrated strong correlation with the annual mean

(rs = 0.91, p < 0.001) and moderate relation with the maximum (rs = 0.64, p = 0.04) and

minimum flows (rs = 0.69, p = 0.02). It displayed significant relations with the pre-

monsoon (rs = 0.86, p = 0.001) and monsoon flows (rs = 0.91, p < 0.001). Similar significant positive relations were present between the maximum wetted width and annual mean (rs = 0.72, p = 0.01), annual maximum (rs = 0.68, p = 0.02), monsoon (rs = 0.81, p <

0.01) and pre-monsoon (rs = 0.6, p = 0.05) flows.

The correlation results between channel width and flow imply a direct relationship between the narrowing pattern of the river and the alterations of downstream flow by the Ranganadi dam. Following dam regulation, downstream flows in Ranganadi underwent significant reductions in magnitude and frequency, including attenuation of flood flows.

66

Chapter 2

Therefore, it can be concluded that the upstream dam had direct impact on the narrowing of the downstream channel by obstructing the water flow. At the same time, the river behaves more erratically even under lower peak flow conditions, since for most part of the year channel morphology adjusts to highly reduced flows (including the base or low flows). Hence, release of large volumes of water from the reservoir during the monsoon season results in flash floods in the downstream reaches where the conveyance capacity of the river had been diminished. The recent floods in July 2017, which once again caused large-scale destruction of floodplain riparian lands and sand deposition, can be considered a recurrence of the 2008 floods. Similar phenomenon of dam-induced reduction in channel capacity, aggradation and channel narrowing resulting in flash floods has been reported by Zahar et.al. (2008), in case of the Medjerda channel impounded by the Sidi Salem dam in Tunisia.

It can be hypothesized that long-term regulation of water and sediment flows by the dam has modified the downstream Ranganadi channel to adapt to a lower flow regime. Besides channel narrowing (specially of the main channel), post-dam abandonment of previously active side channels and a simplification of the channel planform where the river transforms from a multi-thread channel to a single-thread channel can also be considered as direct impacts of flow decrease. The presence of embankments along the river’s length additionally constricts water to pass through a much narrower channel, also making it difficult for the river to deposit the sediment load anywhere else other than the riverbed. In time, the continued in-channel deposition raises the riverbed, making the channel shallower and easier for high flows to overtop the banks. Reduced transport capacity of the channel is particularly problematic during the bankfull and flood flow conditions. These rivers naturally have high contrast between the seasonal flow regimes. Dam regulations exacerbate the contrast in seasonal flow patterns. Therefore, further stress upon the natural pattern by dams only aggravates the situation. This in turn increases the instability of the channel and renders channel changes more unpredictable and complex in nature.

Contrary to the changes in channel width and pattern, the average rates of bank migration have remained relatively similar between the pre-dam and post-dam periods. Differences, however, do exist in the pattern and spatial location of the migration polygons, plus the overall and period specific maximum migration rates. However, these

67

Chapter 2 temporal and spatial differences do not sufficiently establish a direct cause-effect relationship between the dam and bank migration processes. At the same time, migration occurring at specific reaches such as reach 2, can be strongly attributed to flow alterations caused by the dam especially post-2008.

Therefore, the direct impacts of the Ranganadi hydel project can be observed upon the river’s natural flow regime downstream from the dam, while planform features such as braid formation and channel width (overall and reach specific) exhibit the most notable changes to flow alterations. The next chapter examines the hydrological alterations of flow addition in the adjoining Dikrong River and the changes in channel morphology downstream from the powerhouse of the hydel project.

68

Chapter 3

Hydrological and Morphological Changes in River Dikrong

3.1 Introduction

Dams may result in extensive spatial impacts (Braatne et al., 2008), with large geographical spread, over many hundred kilometres on both sides of the obstruction, thereby disturbing the connectivity of the river longitudinally and laterally (Grill et al., 2015). There is a vast body of literature that examines the various downstream impacts on rivers whose flows are commonly reduced from dam closure and regulations, i.e., the flow-deprived river (Ligon et al., 1995; Graf, 2006; Magilligan and Nislow, 2005; Al- Faraz et al., 2014). On the contrary, few studies can be found that discuss the wide-range of impacts experienced by rivers with flow and sediment regime modifications due to reception of additional flows from another river within an inter-basin hydroelectric project (Borgohain et al., 2019). Fluvial and geomorphic responses of a river to elevated flows such as increase in bankfull flow duration or increased annual and monthly discharges would be different from the responses of a river experiencing dam closure and reduction in overall magnitude, timing, and duration of flows (Shields Jr. et al., 2000). Some of the work that has been done on the changes (hydrological and biophysical) exhibited by flow-recipient rivers include Kellerhals et al. (1979) and Church et al. (1995), who reported widening of the river channel in the Lower Kemano River due to reception of diverted flows from the Nechako (Fraser) River (British Columbia) under a hydroelectric project (HEP). In this thesis, River Dikrong is one such flow-recipient river.

The Dikrong is the second river under the influence of the inter-basin RHEP and harbors the powerhouse of the project in Hoj, Papum Pare district, Arunachal Pradesh (AP). It is here that the diverted water from Ranganadi is released after electricity generation. Such water diversion has resulted in the Dikrong transporting more water than its natural capacity across all seasons of channel flow. There is a high probability that the altered flow regime has also affected the geomorphic features of the river.

Therefore, in this chapter I examine the hydrological alterations undergone by the Dikrong following water diversion from Ranganadi and the subsequent spatial and temporal changes in downstream channel morphology.

69

Chapter 3

3.2 Study area - Dikrong basin

Like Ranganadi, the Dikrong is a north-bank tributary of the Brahmaputra River system, originating in the Dafla hills in Arunachal Pradesh at an altitude of approximately 2579 m (Dutt and Datta, 1976). The basin encompasses the hills of Arunachal Pradesh (in its western part), as well as the floodplains of Assam. Located between 27˚00/ and 27˚25/ N latitude, and 93˚00/ and 94˚15/ E longitude, the total catchment area of the basin is 1556 sq.km., out of which 1278 sq.km. lie in Arunachal Pradesh and the rest in Assam (Pandey et al., 2007; Das et al., 2010). Altitude within the basin varies between a minimum of 68 m (in the plains) and maximum of 2879 m (in the hills), while slope varies between 0˚ and 73.17˚ (Fig. 3.1). The total length of the river from its origin in Arunachal Pradesh till its confluence with River Subansiri in downstream Assam is 145 km (Dabral et al., 2007; Bhadra et al., 2011).

Fig. 3.1 Slope variation in the Dikrong basin

The Dikrong flows in a northeasterly direction in the hills for about 48 km, before sharply turning in a southwesterly direction. It enters Assam near the Harmutty Tea estate and runs for another 40 km (approximate) before joining the Brahmaputra. The present Dikrong confluences with River Subansiri instead of a direct confluence with Brahmaputra owing to course shift in the past. The old channel known as ‘Mora Dikrong’ used to join the Brahmaputra about 16 km downstream from its present outfall into

70

Chapter 3

Subansiri (Dutt and Datta, 1976); and the course change likely occurred after the earthquake in 1897 and the old abandoned channel is still present about 8 km to the west of the present Dikrong channel in its lowermost reach (Bhuyan, 2012; Wade, 1800).

Besides its own flow, Dikrong also receives water from a number of tributaries namely, Keyate, Pang Nala, Shu Pabang Nala and Peti Nala on its left bank and Rachi Pabung and Pachin Nadi on its right bank. In the floodplains of Assam, the river receives the Baguli Nadi tributary on its left bank about 21 km upstream of its confluence with Subansiri (SJVN ltd., 2012). Like all other north-bank tributaries of the Brahmaputra, this river too is characterized by a steep gradient with high sediment loads and high rainfall regimes. While the river constitutes a single-channel meandering plan-form in the hills, primarily due to valley confinement, it adopts a braided multi-channel plan-form after entering the downstream floodplains. However, the river again adopts a single-channel meandering pattern near its confluence with the Subansiri. Accordingly, Dikrong can be characterized as a wandering river alternating between meandering and braided channel plan-forms. Like Ranganadi, the riparian floodplains of Dikrong are also prone to frequent flooding during the Northeast monsoon period of May to September. Erosion resulting in loss of agricultural and settlement lands is another recurrent problem faced by the riparian communities in the floodplains. Agriculture is the primary economic activity in the basin with majority of the populace still practicing traditional rice farming.

Rainfall pattern in this basin is similar to the adjacent Ranganadi basin with some variation in magnitude. The basin receives heavy rainfall during the months of June to September like the rest of the Northeast region owing to the South-west monsoon. Around 60-65% of the annual rainfall occurs during the summer monsoon period (Ahmed et al., 2011), while substantial amount of rainfall accompanied by heavy thunderstorm and lightning occurs during the pre-monsoon period of March to May as well (Borgohain et al., 2019). The monsoonal rains might also occasionally extend to the post-monsoon month of October. The summer monsoon, where the amount of rainfall varies from moderate to very high, often result in floods in the low lying plains of Assam. Water level in the river rises and often overtops the banks particularly after periods of incessant and heavy rainfall in the upper catchment. On the other hand, rainfall decreases significantly and is nearly absent during the dry winter period from November to February. The average annual rainfall in the entire Dikrong basin is approximately 3110 mm. Average

71

Chapter 3 annual rainfall in Lakhimpur district (Assam) in the lower catchment varied between 14 mm (December) and 413 mm (July) during the period from 1901 to 2002, while in the upper catchment, average annual rainfall varied between 10 mm (December) and 399 mm (July).

Flow in the Dikrong follows the precipitation patterns of the basin (and the region as a whole) with bankfull discharges and high flows typically occurring during the summer monsoon. The pre-monsoon period also witnesses rise in the water level and the river often reaches bankfull stage before the onset of monsoon. Flow in the river reduces significantly during the dry winter months when there is negligible or no rainfall at all.

Temperature varies distinctly between the higher altitude zone in Arunachal Pradesh and lower floodplains of Assam. Average temperature recorded in Itanagar (Arunachal Pradesh) varied between a maximum of 33.6˚C and minimum of 8.72˚C (SJVN Ltd., 2012). On the other hand, average temperature at North Lakhimpur district in Assam, varied between 28.18˚C maximum in July and 16.88˚ minimum in January during the period from 1901 to 2002.

3.3 Data and methodology

Twenty-eight years of daily flow data from 1987 to 201414 were collected from the Water Resources Department of Assam at Sisapathar G/D station in Lakhimpur District. This data was used to study the hydrological regime of the Dikrong and investigate the flow alterations following completion of RHEP and water transfer. Average annual and monthly discharges that were examined for temporal changes were generated using the Indicators of Hydrologic Alteration (IHA) software (Version 7.1) of The Nature Conservancy, employed similarly in Chapter 2.

Similar to the statistical tests for normality of distribution of the discharge data explained in Chapter 2, the Dikrong hydrological dataset was also subjected to a Kolmogorov-Smirnov test for normality. The data was found to be a non-normal distribution at p < 0.005. Therefore, the hydrological alterations have been evaluated using non-parametric tests in IHA as well as SPSS. Mean values have been used for comparing certain flow conditions, and a Mann-Whitney U test was employed to compare

14 There were no flow data available for the year 1988. 72

Chapter 3

the different categories of flow conditions obtained in IHA between the pre- and post- impact (i.e, pre- and post-dam) periods.

The methodology used for studying the channel changes in Dikrong is similar to the methodology used in Chapter 2. The physical attributes of Dikrong channel and the post-dam alterations in channel geomorphology have once again been determined using remotely sensed data and GIS technology. Landsat (MSS, TM, and ETM) satellite images from 1973 to 2014 were used to detect the channel changes. The 17 individual years that were analyzed are - 1973 (dated 11/15), 1976 (12/14), 1987 (12/03), 1989 (12/08), 1991 (11/12), 1995 (12/09), 2000 (12/06), 2002 (12/12), 2004 (12/17), 205 (11/02), 2006 (12/07), 2008 (11/26), 2009 (11/29), 2010 (12/02), 2011 (10/18), 2013 (12/10), and 2014 (12/29). Channel boundaries and associated channel features were digitized at a scale ranging from 1:10,000 to 1:20,000, based on visual interpretation. Channel widths and bankline were defined using the same definitions as in Chapter 2 for Ranganadi channel, i.e., width was defined in terms of both the bankfull width of the river as well as width of only the wetted channel (i.e., flow transporting channels in the post-monsoon low flow stage).

The reach of the river that has been studied for plan-form changes begins a few kilometres downstream from the powerhouse at Hoj and starts from 27˚10/ E and 93˚46/ N to its confluence with Subansiri near 26˚51/ E and 93˚59/ N latitude and longitudes respectively. Approximate length of the study reach is 59 km. Channel width was measured across 35 sections (D1, D2, D3… D35) separated by equal intervals of 1.5 km along the study reach. Five larger reaches (1, 2, 3, 4, 5) were also demarcated along which the reach specific sinuosity and braiding were evaluated (Fig. 3.2). Reach 1 began from the starting point of the study length till section D8; from D8 to D14 constituted reach 2; from D14 to D22 constituted reach 3; from D22 to D30 constituted reach 4 and reach 5 encompassed sections D30 to D35. All three parameters, i.e., width, sinuosity (sinuosity index, SI) and braiding (braid-channel ratio, B) were calculated using the formulae described in Chapter 2.

73

Chapter 3

Fig. 3.2 Reaches and transects along the river across which channel width, sinuosity and braiding were measured.

Bank migration process was also studied using similar methodology in Chapter 2 in case of the Ranganadi River. Digitized banklines of the concerned study years were overlaid on a GIS platform to generate the areas that have undergone migration because of the lateral movement of the river’s banks. Rate of bankline shifting was calculated using the given equation:

Rmb = (A/Lt1)/Y, where

Rmb is the bank migration (or shift) rate; A is the area of the polygon; Lt1 is the length of the bankline at time t1 (previous year of the two years in comparison) and Y is the number of years between the two years being studied.

Thus, bankline migration was examined across 6 time-frames – 1987-1991, 1991- 1995, 1995-2002, 2002-2006, 2006-2010 and 2010-2014 for periodic changes. The year of dam commissioning (2002) is once again used as a marker to divide the time-frames

74

Chapter 3

into pre-dam (or pre-impact) and post-dam (or post-impact) periods, considering that flow diversion into Dikrong commenced from the given year.

3.4 Results and discussion

3.4.1 Alterations in river hydrology

Precipitation pattern in the Dikrong basin is similar to that of the Ranganadi basin where the highest rainfall occur during June to September and lowest during November to February. Therefore, once again the calendar year (instead of the water year), as was explained in Chapter 2, is used to assess hydrological alterations in Dikrong. These alterations were then compared to the simultaneous temporal changes in channel geometry, assessed from lean season satellite images (November and December imagery).

Flow Duration Curves (FDCs) for the pre- and post-dam period exhibit that the hydrological regime of the river has been largely altered. Post-dam altered flows are higher than the pre-dam natural flows at almost all exceedance probability levels (> 5% to < 95%) (Fig. 3.3). Similarly, the pre- and post-dam hydrographs of median daily flow display the elevation in flow in Dikrong, following water diversion from Ranganadi (Fig. 3.4A).

Figure 3.4B shows the temporal variation in median annual flow from 1987 to 2014, where a marked increase can be observed post-2002. The median annual flow during 1987 to 2002 was recorded to be 50 m3/s (n = 5844), that increased to 119 m3/s (n = 4383, + 138%) after flow addition by RHEP in 2002 with significant statistical difference between the two (p < 0.005). Despite having high fluctuations amongst the post-impact years (from 2003 to 2014), the annual median flow still varied within a higher range of 71 m3/s (lowest in 2003) to 181 m3/s (highest in 2010), against a lower range of flow variations in the pre-impact period. Pre-impact median annual flow ranged between 33 m3/s (lowest in 1989) to 65 m3/s (highest in 1996). Table 3.1 displays the annual median, mean, maximum, and minimum discharge for each year from 1987 to 2014. It can be observed that the lowest annual median flow recorded amongst the post- impact years (71 m3/s in 2003) is still higher than the highest pre-impact annual median flow (65 m3/s in 1996).

75

Chapter 3

Fig. 3.3 Annual flow duration curve obtained from daily flow in Dikrong at Sissapathar gauge station from 1987 to 2014.

(A)

(B)

Fig. 3. 4 Flow regime characteristics of the Dikrong before (1987-2002) and after (2003-2014) dam operation, (A) pre-dam and post-dam hydrographs of daily flow, (B) annual median flow.

76

Chapter 3

Table 3.1: Annual flow statistics of the Dikrong derived from daily discharge data from 1987 to 2014. Year Median Mean Std. Maximum Minimum discharge discharge deviation discharge discharge (m3/s) (m3/s) (m3/s) (m3/s) Pre-dam years 1987 53 106 105 500 12 1989 33 64 60 526 15 1990 57 78 71 696* 15 1991 35 48 39 264 7 1992 48 74 73 736* 20 1993 56 102 112 833* 15 1994 49 71 83 640 15 1995 51 116 140 792* 14 1996 65 113 116 678 14 1997 42 98 114 917** 14 1998 52 123 153 845* 9 1999 55 77 78 452 10 2000 58 90 86 710* 14 2001 41 61 51 301 11 2002 52 75 77 718* 10 Post-dam years 2003 71 90 63 425 13 2004 124 137 81 512 23 2005 175 177 77 625 49 2006 102 124 84 544 21 2007 136 164 105 625 28 2008 131 172 125 957** 40 2009 144 168 120 810* 36 2010 181 201 105 844* 70 2011 106 142 99 916** 52 2012 95 109 59 321 9 2013 96 121 77 499 28 2014 76 125 119 627 15 * = small flood peak, ** = large flood peak (as per IHA statistics)

3.4.1.1 Group 1 alterations

Remarkable increase in median monthly discharge could be observed across all the months from January to December in the post-impact period (Fig. 3.5A). Increase in post- dam discharge was significantly much larger across the low rainfall months of January (+ 173%, Mann-Whitney U test, U = 14, p < 0.001), February (+ 291%, U = 0, p < 0.001),

77

Chapter 3

and December(+ 124%, U = 28, p < 0.005); and also the pre-monsoon months of March (+ 257%, U = 0, p < 0.001), April (+ 319%, U = 2, p < 0.001) and May (+ 192%, U = 21, p < 0.001) (Table 3.2). The change was highest in April at 319%, where it had increased from a median of 29 m3/s pre-regulation to 121 m3/s post-regulation. The wet season months from June to September and post-monsoon months of October and November also exhibited significant increase in flows from pre-dam to post-dam period with percentage of change varying from 32% to 76% (June, + 56%, U = 47, p = 0.02; July, + 32%, U = 43, p = 0.01; August, + 46%, U = 32, p = 0.005; September, + 33%, U = 32, p = 0.005; October, + 50%, U = 28, p < 0.005; November, + 76%, U = 32, p = 0.005).

Thus, it is evident that river flow regime of Dikrong has undergone significant alterations since the transfer of waters from Ranganadi. The annual median flows, however, exhibits a decreasing trend from 2011 onwards though still higher than the pre- dam years (Fig. 3.4B). Table 3.2 presents the alterations in median flow conditions of the IHA parameters from Group 1 to 5 including the percentage of change between the pre- and post-dam periods.

3.4.1.2 Group 2 alterations

All 5 parameters of 1-day, 3-day, 7-day, 30-day and 90-day minimum median flows exhibited significantly high percentage of change (more than 100%) between pre- and post-impact periods at significance, p < 0.005 (Fig. 3.5B, Table 3.2). The highest variation was in the 90-day minimum median flow with 207% increase (U = 1, p < 0.005). However, amongst the five annual maximum flow conditions, only the 90-day maximum median flow showed the highest and significant (U = 45, p = 0.02) change in the post-impact period (+ 49%). The 1-day and 3-day maximum median flow displayed an insignificant decrease post-dam (˗ 10% and ˗ 5% respectively). Thus, amongst the minimum and maximum annual flow conditions, the 90-day extremes displayed the greatest post-impact alteration.

78

Chapter 3

Table 3.2: Difference between the pre-dam and post-dam flow regimes of the Dikrong following completion of the Ranganadi hydel project, obtained from IHA analysis of daily flow data from 1987 to 2014. The asterisks, ‘*’ and ‘**’, indicate significant differences between the two periods at 0.05 and 0.01 confidence levels (Mann Whitney significance test). Pre-dam Post-dam Alteration in median (%) Median Min Max Median Min Max Group 1 – magnitude of monthly flows January (m3/s)** 24 12 32 66 17 168 + 173 February (m3/s)** 19 13 36 74 40 134 + 291 March (m3/s)** 21 11 36 76 39 194 + 257 April (m3/s)** 29 12 58 121 50 210 + 319 May (m3/s)** 59 26 198 172 48 240 + 192 June (m3/s)* 127 38 332 198 96 358 + 56 July (m3/s)* 163 59 405 214 172 318 + 32 August (m3/s)** 130 63 242 189 120 318 + 46 September (m3/s)** 128 65 240 171 95 305 + 33 October (m3/s)** 83 43 139 124 76 190 + 50 November (m3/s)** 45 26 84 79 39 147 + 78 December (m3/s)** 31 14 49 69 21 146 + 124 Group 2 – magnitude of annual extremes 1-day min (m3/s)** 14 7 20 28 9 70 + 100 3-day min (m3/s)** 15 8 20 33 14 71 + 129 7-day min (m3/s)** 15 8 21 35 15 73 + 133 30-day min (m3/s)** 18 11 23 38 17 85 + 115 90-day min (m3/s)** 22 13 32 67 34 121 + 207 1-day max (m3/s) 696 264 917 625 321 957 - 10 3-day max (m3/s) 538 196 775 512 258 641 -5 7-day max (m3/s) 382 145 680 439 230 504 + 15 30-day max (m3/s) 236 106 482 305 188 386 + 30 90-day max (m3/s)* 161 99 324 241 169 320 + 49 Base flow index 0.16 0.09 0.28 0.24 0.13 0.45 + 50 Group 3 – timing of annual extremes Date min 64 25 362 364 5 366 + 36 Date max 187 131 261 205 135 272 + 10 Group 4 – frequency and duration of pulses Low pulse count 8 3 12 0 0 3 - 100 Low pulse duration 5 1 17 2 1 5 - 60 (days)** High pulse count 12 2 21 11.5 5 18 - 4 High pulse duration 3 1 61 4 2 9 + 40 (days) Low Pulse Threshold = 26, High Pulse Threshold = 112 Group 5 – rate and frequency of daily flow changes Rise rate (m3/s/day) 5 3 6 9 6 14 + 70 Fall rate (m3/s/day)** -4 -7 -2 -9 -13 -6 - 141 No. of reversals** 180 109 197 188 162 228 + 4

79

Chapter 3

(A)

(B)

Fig. 3.5 IHA parameters showing the difference between the pre-dam and post-dam flow conditions - (A) Group 1 (median monthly flow conditions), (B) Group 2 (median annual extreme water conditions).

3.4.1.3 Groups 3, 4 and 5 alterations

Amongst the parameters in groups 3, 4 and 5, significant post-dam alteration was displayed by the low pulse duration (- 60%, U = 0.5, p < 0.001), fall rate (- 141%, U = 1, p < 0.001) and number of reversals (+ 4%, U = 2, p < 0.001) (Fig. 3.6). The duration of low pulses changed from a pre-dam value of 5 days to 2 days, while the fall rate decreased from - 4 m3/s/day to - 9 m3/s/day (Table 3.2).

3.4.1.4 Changes in high flows and floods

The IHA separates all high flows in a river into three categories – high flow pulses, small floods, and large floods. High flow pulses indicate all channel flows that are typically bankfull discharges but do not overflow the banks. The default high flow threshold in IHA is constituted by the 75th percentile of all daily flows and all flows greater than the threshold constitute high flow pulses. The high flow minimum, small flood minimum

80

Chapter 3 peak and large flood minimum peak threshold generated in IHA based on daily discharge from 1987 to 2014 for Dikrong is 109 m3/s, 687 m3/s and 866 m3/s respectively.

Fig. 3.6 Group 4 and 5 IHA parameters that showed significant difference in pre- and post-dam water conditions, (A) duration of low pulses (in days), (B) fall rate, (C) no. of reversals. Accordingly, small flood events were observed in 1990 (696 m3/s peak), 1992 (736 m3/s), 1993 (833 m3/s), 1995 (792 m3/s), 1998 (845 m3/s), 2000 (710 m3/s), 2002 (718 m3/s), 2009 (810 m3/s) and 2010 (844 m3/s). Large flood peaks were observed in 1997 (917 m3/s), 2008 (957.4 m3/s) and 2011 (916 m3/s) with the largest flood peak occurring in 2008 (Table 3.1). According to the gauge levels, water level in the river at the gauge station corresponding to each flood peak was higher than the pre-defined danger level of 86.6 m15. However, a comparison of flood flows between the two periods

15 In the Indian system, a river is said to be flooded if its water level at a particular point (usually the gauge/discharge station) is higher than the danger level. Danger levels (DL) of a river that depict the minimum flood thresholds of a river has been pre-determined by the Central Water Commission, New

81

Chapter 3

reveals that there has been no significant change in the median flood peak (small and large). The pre-impact median small flood peak was 736 m3/s (mean = 757 ± 56, n = 7) that changed to 827 m3/s (mean = 827 ± 24, n = 2) post-impact, marking an increase of 12%. When median is calculated for all the small and large flood peaks together, then pre-dam median flood peak is 764 m3/s and post-dam median flood peak is 880 m3/s, marking an increase of 15%.

3.4.2 Changes in river morphology

3.4.2.1 Channel sinuosity and braiding

As is common with rivers passing through a hilly terrain confined by valley walls, the Dikrong adopts a meandering pattern in the hills of Arunachal Pradesh. The river modifies into a multi-stream braided channel as it flows into the flatter floodplains of Assam. Accordingly, at reach 1 (within the study length of the river) the sinuosity is high (SI ~ 2), while the braid-channel ratio is negligible (B ~ 1.2). Reaches 2 and 3 exhibit a highly braided multi-channel planform, especially from 1987 onwards. However, towards the lower reaches, i.e., reach 5 where the river approaches its confluence with the Subansiri, the planform again becomes sinuous with SI = 1.45 and B = 1 (as in 1973). Table 3.3 displays the temporal variations in the overall and reach-specific sinuosity and braiding of the Dikrong channel.

The overall sinuosity decreased steadily from 1.87 in 1973 to 1.66 in 2002, and has since stabilized. This decrease could be attributed to shortening of the channel length owing to elimination and straightening of meanders through neck and chute cut-offs. Braiding, on the other hand, increased sharply from 1.12 in 1973 to 1.55 in 1987 and 1.86 in 1991. From 2002 onwards, braiding displayed a steady increase from 1.68 to 1.86 in 2010 and 1.81 in 2014. Sinousity and braiding tend to correlate negatively (Friend and Sinha, 1993), which is found to be consistent with the contrasting pattern of change displayed by the two indices (SI and B) in Dikrong. This also marks the river’s transition from a primarily single-channel meandering plan-form in 1973 to one that is increasingly braided and multi-channeled in 2014.

Delhi, along with different State Government Engineers and fixed at important G/D sites for majority of the Indian rivers. Accordingly, a water level equal to or higher than the DL indicates that the river is flowing at flood conditions at that site (Dhar and Nandagi, 2000).

82

Chapter 3

Table 3.3: Temporal and spatial variations in the sinuosity indices (SI) and braid- channel ratios (B) measured across the 5 reaches of the Dikrong as well as overall examined length from 1973 and 2014.

Years Overall Reach specific

1 5 2 3 4 (starting (D30 till (D8 to (D14 to (D22 to point to confluence D14) D22) D30) D8) point) SI B SI B SI B SI B SI B SI B

1973 1.87 1.12 2.28 1 1.55 1 1.28 1.38 1.51 1.19 1.45 1 1987 1.90 1.55 2.22 1.20 1.25 1.95 1.19 2.17 1.55 1.48 1.88 1.15 1991 1.81 1.86 2.08 1.08 1.20 3.35 1.13 2.63 1.19 1.84 1.93 1.24 1995 1.68 1.46 2.10 1.14 1.22 2.32 1.13 1.84 1.23 1.35 1.47 1.02 2002 1.66 1.68 1.99 1.26 1.28 1.96 1.56 2.59 1.19 1.61 1.47 1.03 2004 1.61 1.78 2.00 1.19 1.28 2.44 1.14 2.10 1.15 2.24 1.25 1.25

2006 1.58 1.70 1.97 1.11 1.28 2.18 1.19 2.03 1.23 1.70 1.32 1.22

2008 1.65 1.60 1.99 1.09 1.30 2.63 1.18 2.13 1.18 1.07 1.36 1.16

2010 1.64 1.86 2.06 1.00 1.27 1.94 1.20 2.50 1.15 2.30 1.32 1.72

2014 1.64 1.81 1.97 1.05 1.34 1.92 1.26 2.47 1.16 2.25 1.35 1.42

Secondary channel formation induced by the avulsions in 1991 and formation of mid-channel bars were responsible for the high B in 1991 (1.86) as opposed to. The same planform modifications resulted in marked widening of the 1991 channel, discussed later in the chapter (in section 3.4.2.2). However, some of the newly formed bars and side channels were abandoned in the years that followed, thereby lowering B in 1995 as well as 2002.

Similar to the temporal change in overall SI and B between 1973 and 1987, reach- specific sinuosity and braiding also underwent major changes between the two years, particularly in reaches 2, 3, 4 and 5. Development of multiple threads around channel (braid) bars from section D9 to D30 (Fig. 3.7A) increased B from 1 to 1.95 in reach 2, from 1.38 to 2.17 in reach 3 and from 1.19 to 1.48 in reach 4. With respect to sinuosity,

83

Chapter 3

(A)

(B)

Fig. 3.7 Transformations in channel planform in Dikrong between 1973, 1987 and 2014, (A) increased development of

braid bars and multiple channels along reaches 2, 3 and 4 in 1987 contrary to a predominant single- channel meandering planform in 1973, (B) continued increase in braiding and elimination of meanders in reaches 4 and 5 from 1987 to 2014.

reaches 1, 3 and 4 did not exhibit much change from 1973 to 1987, but decreased in reach 2 and increased sharply in reach 5 (1.45 to 1.88). Reach 5 remained primarily meandering in pattern through 1987 to 1991 (SI = 1.93), but became sinuous by 2002 (SI = 1.47) and 2014 (SI = 1.35) (Table 3.3). Simultaneously, braid-channel ratio at reach 5 increased

84

Chapter 3

from 1.15 in 1987 to 1.72 in 2010 and 1.42 in 2014. Similar increase in B was also observed in reach 4 (Fig. 3.7B). This change suggests a downstream progression in braiding and its expansion over a longer stretch of the river, primarily post-2008.

In addition to higher discharges, the amount and changes in bed-load sediment and other factors such as slope, nature of bed and bank material also influence braiding intensity (Friend and Sinha, 1993). The Dikrong transports heavy sediment loads from the upper catchment, which are deposited over a shallow river-bed in the flatter floodplains (SJVN Ltd., 2012; Borgohain et al., 2019). According to the Detailed Project Report for the Doimukh Hydroelectric project prepared by SJVN Ltd. (2012), the average sediment load (including suspended and bed loads) during 1976 to 1991 was 1027.19 ha m. Hence, a comparative study of changes in its sediment load between pre- and post-dam periods would help understand RHEP’s role in increasing channel braiding in the post-dam period. However, analysis of sediment flow pertaining to Dikrong was beyond the scope of this study owing to insufficient data. In addition, modification in channel sinuosity and braiding in Dikrong had begun pre-2002, i.e., before addition of flows by RHEP. Therefore, it cannot be concluded with certainty whether the continued increase in braiding during the post-dam years could be attributed solely to the addition of flows by the hydel project, except in the form of augmenting the natural process of change.

3.4.2.2 Changes in channel width

The Dikrong widens considerably at its middle segment in the plains where the planform is largely braided, i.e., reaches 3 and 4 (from section D14 to D30) and is comparatively narrow at its meandering reaches (within the hills and near the confluence with Subansiri), i.e., reaches 1 (from D1 to D8) and 5 (from D30 to D35). Width in reach 3 varied between 574 m in 1987 and 870 m in 2000; reach 4 varied between 469 m in 1973 and 974 m in 1991. Width in reaches 1 and 5 varied between a lower range of 209 m in 1973 and 273 m in 1976 for the former and 310 m in 1973 and 506 m in 2005 for the latter. Reach 2 was observed to be intermediate with widths varying between 530 m in 1973 and 648 m in 1991. In the post-dam period, distinct increase in channel width was found in the lower reaches of the river, i.e., 4 and 5. Figure 3.8 shows that width in reach 4 increased sharply in 1991 and again in 2008, becoming the widest segment of the river thereafter. Similar rise in width was observed in reach 5 immediately post-2002.

85

Chapter 3

Fig. 3.8 Temporal variation in the average bankfull channel width along the five reaches from 1973 to 2014.

Table 3.4: Average, maximum and minimum bankfull and wetted widths of River Dikrong from 1973 to 2014 obtained from 35 sections along the study length of the river. Year Bankfull width Wetted width

Average Stdev Max Min Average Stdev Max Min (m) (m) (m) (m) (m) (m) 1973 421 215 1084 63 206 94 479 63 1976 425 202 937 75 179 110 451 48 1987 392 246 886 54 155 73 375 54 1989 386 239 865 58 167 88 432 56 1991 567 386 1506 53 192 110 404 53 1995 462 268 1049 56 187 98 423 56 2000 547 354 1216 97 159 95 495 38 2002 525 371 1266 60 178 93 394 53 2004 539 327 1359 63 216 112 423 63 2005 531 332 1349 85 222 112 464 60 2006 462 294 1393 59 158 77 393 59 2008 573 403 1774 66 227 96 423 66 2009 498 374 1772 66 207 92 460 66 2010 476 381 1790 53 207 90 381 53 2011 508 383 1821 69 239 132 543 37 2013 518 421 1922 42 219 121 591 42 2014 536 404 1949 57 243 135 633 57

Temporally, the overall bankfull channel width (n = 35) varied between an average of 386 m (± 239, 1989) and 573 m (± 403, 2008) from 1973 to 2014 (Table 3.4). Observed variations showed a substantial widening pattern of the river post-1990 till 2014 (Fig. 3.9A). Therefore, width increase can be traced back to the pre-dam period, which

86

Chapter 3

suggests that the post-dam increase in average channel width is not solely dam induced. Nevertheless, comparison between the average pre-dam and post-dam widths reveal that the bankfull channel was 10% wider during the latter period. During the pre-dam period of 1973 to 2002, the channel exhibited an average width of 466 m (± 72, n = 8), while during the post-dam period of 2002 to 2014, the average was 516 m (± 34, n = 9). This suggests that even though flow augmentation by the RHEP may not be a direct or the only cause of channel widening in Dikrong, yet its contribution to the width changes cannot be discounted.

(A)

(B)

Fig. 3.9 Temporal variation in the overall channel widths in Dikrong measured over 33 transects along the river’s study length, (A) bankfull channel width, (B) wetted channel width.

There were two distinct points of sharp increases in average bankfull width – 1991(567 m) and 2008 (573 m). Channel widening in 1991 was driven by the localized width increases (greater than 50% increase) across sections D1, D5, D13, D21, D23, D25, D27, and D29. The given sections showed different types of plani-metric modifications between the 1989 and 1991, because of which the width change was so high. While D1 and D5 showed a normal widening of the wetted channel, sections D13, D21, and D29

87

Chapter 3

changed from a single-channel pattern to a braided pattern with formation of mid-channel bars, thus leading to increased widths. At section D27, in addition to a general widening of the primary channel, channel avulsion had occurred in 1991, that led to the formation of a new secondary channel on the right bank. The newly formed side channel cut across the meander bend and enclosed the adjoining lands into a vegetated island bar, thus increasing the total width of the river across D27. Figure 3.10A displays the various planform modifications in Dikrong in 1991.

The channel section at D23 and D25 exhibited interesting planform changes that caused the sectional widths to increase sharply in 1991 as well as again in 2008 (Fig. 3.10B). These changes were of a transient nature where the modifications that appeared in 1991 disappeared for a while (between 2004 and 2006), only to resurface in 2008. Initially in 1991, a chute cut-off resulted in avulsion of the primary channel and straightening of the meander curve. The old channel was retained for some years as a partially active channel and hence the high sectional bank-to-bank width. Between 2004 and 2006, the side channel appears fragmented and abandoned, but was formed again (with some shift in position) in 2008 leading to an increased width.

Bankfull width increase in 2008 was again found to be a result of planform changes and localized width increases at sections D1, D5, D13, D17, D21, D25 and D33. Channel widening at D17 was initiated in 2000 itself, which had also increased the average bankfull width during that year.

Therefore, it was observed that specific sections along the river’s length exhibited higher widening in both wetted and bankfull widths than a uniform increase in width across all sections. The channel at section D17 exhibited the maximum width through all the years from 2000 to 2014 while displaying a continuous rise each year. Width at the section initially increased due to the formation of a secondary channel along the river’s right bank in 2000 (between D16 and D19) (Fig. 3.10C). The area in between the newly formed channel and the primary channel of the river was enclosed into a vegetated mid- channel island. Over the years, as the two channels meandered and shifted apart through lateral erosion along the banks, the size of the island bar increased along with the increase in the sectional width (Borgohain et al., 2019). Mid-channel bars, in turn, reduce the cross-sectional area of the river at the point of its formation and consequently the water channels cut into the banks laterally to accommodate the discharge (Bora, 2004).

88

Chapter 3

(A) (B) (C)

Fig. 3.10 Planform changes that caused width increases during different years, (A) increase in sectional widths between D21 and D29 in 1991 due to avulsions and secondary channel formation, noticeably at D27, (B) changes at sections D23 and D25 in 2008 that were a repetition of past changes in 1991, (C) initiation of a secondary channel on the right bank between D15 and D19 that grew into two prominent channels running near parallel but bending away from each other by 2014, thereby increasing the sectional width at D17.

Widening of the river at section D17 (wherein width was highest in 2014 at 1949 m) increased the overall channel width considerably, especially in the post-dam period.

89

Chapter 3

The associated bank erosions on both sides of the river have also severely affected the people residing by the banks through loss of agricultural and settled lands.

In case of the channel width occupied by the wetted area of the Dikrong, average width across the 18 studied years varied between 155 m (± 73, n = 35) in 1987 and 243 m (± 135, n = 35) in 2014 (Fig. 3.9B). The average pre-dam wetted width was calculated to be 178 m (± 17, n = 8) while post-dam average wetted width was 215 m (± 25, n = 9). Thus, the wetted area of the river also displayed a widening pattern with 20% increase. Increase in the wetted width was higher than the increase in bankfull width during the post-dam period.

Later in section 3.4.4, Spearman rank correlation analyses between the different flow variables and channel parameters have been presented. Based on the rho (rs) and p- values, maximum channel width showed strong correlation with the annual mean and total discharge. This suggests that the continued temporal increase of the maximum width at section D17, especially from 2008 onwards, is influenced by the post-dam flow increases in the river by the RHEP.

3.4.3 Bankline migration, erosion, and deposition patterns

Shifts in the position of channel banklines were studied in terms of both periodic changes as well as overall shift between 1987 and 2002, which corresponds to 15 years of unaltered pre-dam period; and between 2002 and 2014 corresponding to 12 years of altered post-dam period. Periodic bankline migration was analyzed across six time- periods from 1987 to 2014.

3.4.3.1 Overall bank migration between 1987-2002 and 2002-2014

Dikrong exhibited a wide range of plan-form adjustments both spatially and temporally throughout 1987 to 2014, with occurrences of meander neck cut-offs and straightening of meanders at varied spatial locations along the downstream length of the river. The overall shifting of both banklines were found to be slightly higher during the post-dam period of

2002-2014 at 6 m/y (RB) and 7 m/y (LB) average shift rates (Rmb). At the same time between the two banks, the left bank exhibited more intense shifting than the right bank

during both periods of study. The maximum migration rates (max Rmb) were also evaluated to be higher for both banks during the post-dam period. Maximum rate of

90

Chapter 3

migration by the right bank during 2002-2014 was 29 m/y (polygon across D11 and D13, aggradation along the right riparian plain) and 36 m/y by the left bank (polygon across D25 and D27, erosion along the left riparian plain). Table 3.5 displays the difference in overall bankline migration, erosion, and deposition patterns between the pre-dam and post-dam time-frames. Most of the post-dam changes in bankline position were a continuation from the pre-dam period. Therefore, it cannot be stated with certainty that the flow augmentation by the upstream dam initiated the shifting and eroding patterns in the river. These processes have been continuing from pre-dam times; but the daily fluctuations in water levels has most likely exacerbated the problem of loosened bank materials and increased bank instability. Borgohain et al. (2019) also stated that the absence of large woody trees along the banks of the Dikrong (for most part of its length in the floodplains), resulted in the banks being vulnerable to erosive forces.

Table 3.5: Overall bankline migration, erosion, and deposition of Dikrong between pre- dam (1987-2002) and post-dam (2002-2014) time-periods.

1987 – 2002 2002 - 2014 Right bank Left bank Right bank Left bank Average rate of bankline 5.3 5.9 6.1 7.3 migration (Rmb) (m/y) Stdev 6.1 5.5 6.4 7.9

Maximum Rmb (m/y) 27 22.2 29.3 36.1

Minimum Rmb 0.5 0.6 0.8 1.2

Total area of erosion ( km2) 6.5 2.9 3.4 4.8

Annual rate of erosion ( km2/y) 0.4 0.2 0.3 0.4

Total area of deposition (km2) 1.4 5 3.2 4.7

Annual rate of deposition 0.1 0.3 0.3 0.4

* Rmb = rate of bankline migration, Stdev = standard deviation

Finally, three processes could be distinguished that explain the changes in the position of banklines in Dikrong during both pre-dam and post-dam periods - i) Lateral migration of the banks in opposite directions with the two water channels flowing adjacent to the banks resulting in erosion of the adjoining lands. This was the

91

Chapter 3 primary process of erosion witnessed across sections D16 and D17 where the river has been growing progressively wider, particularly post-2002 (see Fig. 3.10C). During field survey on, 2014, it was witnessed that the water channels pressed against the banks near D17, resulting in rapid erosion along the stretch. Substantial loss of land and vegetation was observed due to rapid bank failure (slumping of banks) within a short span. The photographic images in Fig. 3.10 were taken on June 3, 2014, where the road and a few betel nut trees could be seen standing, along with the evidences of earlier erosion. However, after 10 days, the entire road visible in the image and the trees were eroded away. Locals believed that frequent and fast fluctuations in the water level of the river owing to water release from the RHEP powerhouse at Hoj, loosens the bank material causing bank failure. The satellite image in Fig. 3.11 dates December 29, 2014 and one

Fig. 3.11 Erosion along the left bank due to the lateral movement of the banklines in opposite direction across section D17, where the river had split into two prominent near equal sized and parallel channels. The images in the right were taken during a field visit to the site in June 2014. can see that the location, at which the photographic image was obtained in June earlier, lies well inside the water channel. Bank failure (slumping) during the falling stages of river discharge has been reported commonly in the Brahmaputra River (Goswami, 1985). Bora (2004) stated that when water levels in a river during the flood stage rise, the water seeping into the banks (primarily composed of weakly held sand materials) increases the pore-pressure of the bank wall. As the water goes down during the falling stages, a

92

Chapter 3

reverse flow of sand and silt into the river occurs that results in ‘subaqueous bank failure’ (Bora, 2004). Based on the key observations put forward by the local correspondents and personal observation, a similar explanation is applicable to the bank erosion mechanism of Dikrong. It is possible that the daily fluctuations in water level in the river following flow addition by RHEP have further aided the rate of erosion.

(ii) Lateral migration of the active channel belt as a whole that resulted in gradual erosion along one bank and aggradations along the other. The reach from D19 to D25 exhibits this movement (Fig. 3.12).

Fig. 3.12 Lateral westward migration of the active channel belt from D19 to D25 causing erosion along the right bank and aggradation along the left.

(iii) Rapid channel diversion of the primary channel or avulsion led side channel formation and meander chute or neck cut-offs (reach between D33 and D35) that reconstructed the bank margins across time and space.

Considerable spatial variation existed in the location of the migration polygons with large shift areas and rates before and after dam completion. Explained below is the

93

Chapter 3

migration pattern of the two banks before and after the intervention of RHEP and the spatial distribution of the migration polygons with the most notable shifts.

During the pre-dam period, lateral migration of the right bank could be witnessed in the reach from D19 to D25, where the active belt of the river (as a whole) had shifted 2 to the west over an area of 3.7 km (Rmb = 25 m/y, the second largest migration by the right bank between 1987 and 2002) (Fig. 3.12). At the same time, maximum Rmb of 27 m/y was caused by an avulsion of the primary channel and resultant meander neck cut-off between D33 and D35 that had also resulted in the lowering of the total and localized channel sinuosity (Fig. 3.13). As time passed, the channel at the said sections adopted a meandering pattern where the meander curvature has only been increasing over the years. Hence, the possibility of a chute cut-off in the future is highly likely. Contrarily, the polygon with maximum migration rate during the post-dam period (29 m/y) was located across D11 and D13, where an eastward shift of the active channel belt resulted in deposition along the right bank and consequently erosion along the left bank. In general, the post-dam period witnessed higher deposition with the emergence of a total floodplain area of 3.2 km2, against 1.4 km2 in the pre-dam period.

Fig. 3.13 Meander neck cut-off between D34 and D35 that shifted the channel to the east.

94

Chapter 3

Erosion was observed to be quite high along the right riparian plain (especially in the pre-dam period) compared to the left bank plains. Between 1987 and 2002, 6.5 km2 of total area along the right bank had been eroded away through a lateral westward shift of the bankline, compared to 3.4 km2 between 2002 and 2014. The maximum pre-dam erosion occurred between sections D19 and D25 at 3.7 km2 of sectional eroded area. Field visit to the concerned site and interviews with the riparian communities revealed that severe erosion by the river had destroyed personal and public property, grazing, and flourishing agricultural lands. School buildings, public health centers etc., roads, and embankment had all been eroded away as the river shifted west. The erosion process still continues and the re-constructed embankment and other bank protection measures have not been able to arrest the progress so far. Between 2002 and 2014, an additional 2 km2 of riparian area was eroded, with a spread to sections D16, D17 and D18. Many residents complained that the daily fluctuations in the water level of the river caused by the upstream hydel project has further aggravated the erosion problem and loosened the bank materials.

In case of migration by the left bank, the channel from D19 to D27, and at D33 and D35, exhibited notable shifts during the pre-dam period. The westward shift from D19 to D27 was in tune with the westward movement of the right bank explained previously. The left bank shift at D33 and D35 is also the chute cut-off that has been discussed in the right bank migration pattern. However, the post-dam period saw the highest migration rate of 36 m/y in the left bankline between sections D24 and D28, which also recorded the maximum area of erosion at 2.5 km2 (Fig. 3.14). The primary channel had shifted eastward, encroaching into the left floodplain, and progressively eroding the lands over the years. The right bank however did not exhibit a simultaneous eastward shift since the right bank margin was formed by a side channel, which was actually a vestigial channel of the original meander bend. Width of the river at section D25 also increased consequently, reaching 1246 m in 2014. The overall erosion by the left bank was found to be higher during the post-dam period with 4.8 km2 of total eroded area against 2.9 km2 of pre-dam eroded area; whereas erosion along the right bank was higher during the pre-dam period.

95

Chapter 3

Fig. 3.14 Migration between sections D24 and D28 that recorded the highest rate of bankline migration during the post-dam period (2002-2014) at 36 m/y (left bank) and the maximum area of 2.5 km2 (shaded area marked as ‘e’). The figure shows how the primary channel had shifted eastward eroding along the left floodplain and reconstructing the left bankline in the east, while the right bankline persisted due to the formation of a side channel.

3.4.3.2 Periodic bankline migration

The temporal comparison of migration rates between the overall pre-dam (1987-2002) and post-dam (2002-2014) periods provide an understanding of whether or not bank migration was affected by flow addition following water diversion by RHEP. The periodic rates, on the other hand, indicate when within the overall time-period did specific shifts take place. Within the pre-dam periods, highest average migration rate by both banks occurred during 1991-1995 at 18 m/y (RB) and 16 m/y (LB). Migration amongst the post-dam periods was highest during the period immediately following flow addition by the hydel project, i.e., 2002-2006 at 15 m/y (RB) and 19 m/y (LB). Table 3.6 further details the various episodic rates of bankline shifting and corresponding area of erosion and deposition that had taken place on both banks of the Dikrong from 1987 to 2014.

96

Chapter 3

3 13 14 96 0.8 0.2 LB 1.48 0.37 point) below D35 below D35 (confluence (confluence 2014 (Y=4)2014

4 16 20 2010 - RB 120 1.55 0.39 1.82 0.46 point) below D35 below D35 (confluence (confluence

2 4 1 to

14 18 85 1.3 LB D23 D23 D27 0.33 2010 2010

(Y=4) 3 14 12 55 2006 - 1.4 1.8 RB and D11 D11 D12 0.35 0.45

3 to

19 18 83 4.4 1.1 2.2 LB D21 D21 D24 0.55 2006 2006

(Y=4) 3 15 12 57 2002 - RB D35 1.91 0.48 1.41 0.35

9 2

10 38 LB and D19 D19 D20 3.33 0.48 3.73 0.53 2002

(Y=7) 9 8 2 to 32 1995 - 1.9 RB D17 D17 D22 3.94 0.56 0.27

3 16 23 1.2 LB 165 D27 3.28 0.82 4.79

1995 (Y=4)1995 3 18 19 94 RB D35 1.93 0.4 8 3.79 0.95 RB = right bank, LB = left bank = rightleft bank, LB RB , D34 and D34 1991 -

2 15 17 85 LB D23 4.29 1.07 1.04 0.26 igration, erosion and deposition patterns of the Dikrong between 1987 and 2014 along with the most affected most affected 2014 along the with 1987 and Dikrong thebetween of patterns deposition erosion and igration,

1991 (Y=4)1991 3 14 16 RB 103 D25 4.34 1.09 1.09 0.27 D22 to D22 1987 -

(m/y) (m/y) (m/y) (m/y) mb mb mb mb

) /y) ) /y) 2 2 2 2 Periodic bankline m bankline Table 3.6: Periodic channel sections. Stdev R Maximum Section(s) exhibiting Section(s) exhibiting R maximum R Average deviation standard = Stdev migration, of rate = Rmb Total deposited area area Total deposited (km erosion erosion of Rate (km deposition Rate of (km Minimum R Minimum Total Total area eroded (km

97

Chapter 3

Migration along specific segments of the river displayed remarkable variations, as could be observed from the maximum migration rates amongst different periods. The highest maximum migration rate displayed by the right bankline was recorded at 120 m/y

during 2010-2014 (average Rmb = 16 m/y). Maximum Rmb by the left bank was highest during 1991-1995 at 165 m/y (average Rmb = 16 m/y). The high fluctuation (temporal and spatial) in maximum migration rates is quite contrary to the low variation in the average migration rate.

Average Rmb of the right bank was lowest during 1995-2002 at 9 m/y and highest

at 18 m/y during the preceding period of 1991-1995. Similarly, average Rmb of the left bank was lowest at 10 m/y during 1995-2002 and highest at 19 m/y during 2002-2006. Each time-period also displayed a wide spread between the minimum and maximum migration rates. This probably implies that bank migration was localized and concentrated in certain sections of the river, while the rest of the channel segments remained relatively stable and resilient. Such spatially specific change pattern was also witnessed in the bankfull channel width modifications and segment-specific braiding patterns.

The river segments where the polygons with high migration rates and areas were located are likely the more vulnerable and unstable reaches. Maximum Rmb during 1987-

1991 (103 m/y RB Rmb and 85 m/y LB Rmb) occurred from D22 to D25 (Table 3.6). The shift had resulted from channel avulsion and meander straightening that caused the primary channel to shift west with erosion of the right floodplain. At the same time, the old channel persisted as a small side channel16. Polygon with maximum migration in the following period, i.e., 1991-1995, was located at D34 and D35 for the right bank17 (max

Rmb = 94 m/y) and at D27 for the left bank (max Rmb = 165 m/y, lateral westward shift of the primary channel resulting in erosion along the right bank and aggradation along the

left bank). 1995-2002 polygon with maximum Rmb was located across D17 to D22 for the

right bank (max Rmb = 32 m/y) and across D19 and D20 for the left bank (max Rmb = 38 m/y). 2002-2006 maximum migration rate polygon occurred across D35 along the right

bank (max Rmb = 57 m/y) and across D21 to D24 along the left bank (max Rmb = 79 m/y).

16 See Fig. 3.9A displaying the planform changes of 1991 Dikrong channel that had increased the bankfull

channel width. The same changes resulted in maximum Rmb between 1987 and 1991. 17 The meander neck cut-off across sections D34 and D35 obtained during the overall pre-dam period (1987-2002) and shown in Fig. 3.11 had actually occurred between 1991 and 1995.

98

Chapter 3

Fig. 3.15 Centerlines of the Dikrong channel for the years 2002, 2006, 2010 and 2014 and right banklines of the Subansiri during the same years showing the pattern of frequent shifts in the point of confluence between the two rivers.

The position of the confluence between the Dikrong and Subansiri shifted frequently shifting during different years. It was at this location that the maximum migration rates of 120 m/y (RB) and 96 m/y (LB) were observed in 2010-2014. The shift in the position of the confluence was, however, influenced by the channel-shifting pattern of both rivers, i.e., shifting of Subansiri as well. The confluence shifted by more than 2 km to the east between 2002 and 2006 as the Subansiri River eroded along its right bank (encroaching into the floodplain caught between the two rivers). However, by 2010 the confluence again shifted ~ 1 km to the west and another 1 km by 2014, reverting back to

99

Chapter 3

near the old confluence point of 2002. Figure 3.15 shows the various shifts in the position of the confluence from 2002 to 2014.

The spatial comparison of overall and periodic migration pattern indicates that the floodplain reach of the Dikrong in its lower catchment in Assam (from D15 to D30) has been regularly exposed to intense bankline shifting by unstable banks and progressive erosion. Some of the riparian villages that are located within the mentioned sections and had experienced recurrent bankline migration were No. 80 Solmari, No. 65/68 F.C. Grant, Bango Gaon, Mornoi Gaon, Bihpuria, Keyamora etc. on the right bank and Merbeel, No. 79/82 Grant, Gereluwa, Kalabil Mornoi Nepali and Chicha Pathar etc. on the left bank. Like most other riparian lands, the banks of Dikrong and the adjacent floodplain are also put to agricultural use, with paddy being the chief cultivated crop. Recurrent bankline shifting and the resultant erosion and deposition processes pose a major threat to the use of the fertile riparian land for cultivation purposes, especially where agriculture is the major source of livelihood. Loss of land through erosion and sand casting are the two foremost detrimental effects of persistent bankline shifting on riparian populations.

3.4.4 Comparison between flow conditions and channel width

An active river system constantly changes to achieve a state of equilibrium, and complex overlapping of multiple factors besides water flow can initiate the changes. It is, therefore, difficult to attribute flow alone as a strict, or the only cause of planform changes, particularly when there can also be other contributing factors such as sediment discharge18. Nevertheless, Spearman’s rank correlation coefficient analyses were run between average and maximum bankfull widths with different flow conditions to understand their association.

The mean monthly flows were divided into three categories based on the precipitation patterns in a calendar year – pre-monsoon mean flow (comprising of the mean monthly flows from March to May), monsoon mean flow (mean monthly flows from June to September), and the post-monsoon mean (mean monthly flows for the two dry months of November and December). These flow categories were then compared with the average and maximum channel widths for Dikrong. Comparisons were also

18 However, it has already been mentioned earlier that sediment analysis could not be done for the Dikrong (or the Ranganadi) due to lack of adequate data.

100

Chapter 3

drawn between width and the annual mean and maximum flow during each study year. Since no flow data were available prior to 1987, the width values for 1973 and 1976 have been excluded from the correlation analyses.

The average bankfull width did not exhibit a correlation with any category of river flow. It is possible that channel width did not directly correspond to the changes in river flow. The maximum bankfull width, however, exhibited significantly moderate relation with the annual mean flow (r = 0.53, p = 0.04, n = 15) and the pre-monsoon mean flow (r = 0.53, p =0.04, n = 15). This pattern of relation between maximum width and flow, absent in case of the average bankfull width, suggests that width changes in response to flow alterations in the Dikrong are not uniformly spread along its entire downstream length. Instead, specific sections undergo changes while the rest of the channel remains comparatively stable, implicating that flow elevation in the post-dam period influenced width changes along certain sections, but not the entire downstream reach.

The maximum bankfull width did not exhibit a correlation with the annual maximum and monsoon mean flow (r = 0.24, p = 0.38, n =15 and r = 0.3, p = 0.27, n = 15 respectively). This conforms to the observations by Kellerhals et al. (1979), where the Kemano River displayed channel widening and straightening in response to increased mean flows even though large floods had not increased post-diversion. Brandt (2000) had also asserted that channel changes can occur due to alterations in ‘intermediate flows’ as well and bankfull flows should not be the sole focus. Next, I compared the flood peaks (small and large floods) with the average and maximum width. Once again, it was the maximum bankfull width rather than the average that related strongly to the annual flood peaks with r = 0.79, p = 0.04 (n = 7). A direct increase in the average bankfull width could be witnessed only in 2008 where it had increased to 573 m, also the highest amongst all the study years. 2008 also recorded a flood peak of 957 m3/s, which was the highest amongst all recorded flood peaks from 1987 to 2014. This suggests that not all bankfull flows resulted in width adjustments in the river, rather those that culminated in floods (Borgohain et al., 2019).

With respect to the wetted channel, average wetted width correlated significantly with the flood peaks (r = 0.96, p = 0.001, n = 7), annual mean (r = 0.6, p = 0.02, n = 15), pre-monsoon (r = 0.61, p = 0.02) and post-monsoon flow (r = 0.6, p = 0.015, n =15). No relation was found between the maximum wetted width and downstream flow.

101

Chapter 3

There exists a probability that the response of channel width to flow would become more unpredictable in the future and spread to multiple sections along the river’s length. This can be hypothesized keeping in view the cascade of small hydel projects planned upon the river both upstream and downstream from the powerhouse at Hoj, one of which is near completion. Although, labeled as RoR schemes, three projects namely the 60 MW Turu HEP, 52 MW Doimukh HEP and 110 MW Pare HEP include barrages and dams for water storage, and downstream flow will be subject to regulations as per the electricity generation schedules. The cumulative effect of these projects would significantly modify downstream flow in Dikrong in the future and impact upon the river’s morphology downstream from these structures. The patterns of bankline migration and erosion in Dikrong were observed to be localized and episodic, indicating no clear linkage to the flow changes between the years. Channel avulsions and lateral movement resulting in erosion of the adjacent floodplain was evident both before and after flow addition by RHEP.

One of the possible reasons for the absence of a direct and uniform change in channel geometry over the entire length of the river according to the changes in high flows could be ‘bank-stabilizing structures such as embankments’ (Borgohain et al., 2019). Embankments as a flood control measure are a common sight in the study area. Yet, there are sections where the structures have been eroded away thereby exposing the adjacent floodplain. Embankments are often unsuccessful in containing large floods that can easily breach a weakened and failing structure (Goswami, 2008, Bhattachaiyya and Bora, 1997). They are ‘partially effective’ in withstanding frequent floods and are unable to resist erosion in the long-term, which finally results in changes in channel geometry (Northwest Hydraulics Consultants, 2006). At the same time, these structures facilitate aggradation of the river-bed, by constricting water, silt, and sediment flow to a narrow channel. The north-bank tributaries of the Brahmaputra exhibit shallow river-beds because of the deposition of large amounts of sediment loads carried down from the upper catchment; that naturally makes flooding and bank overtopping easier with or without embankments. Consequently, large floods such as the ones in 1991 and 2008 may result in overtopping and avulsions, thereby affecting bankline migration.

From the aforementioned observations and analysis it is quite evident that the downstream flow pattern of the Dikrong has been significantly modified following

102

Chapter 3

addition of Ranganadi flows in 2002. However, the channel adjustments shown by Dikrong over the same time-scale are erratic (such as bankline migration) and marked by fluctuations between the various periods. The changes in one period are spatially distinct from another period, which makes it difficult to generalize the trend of change and establish a strong causality with the flow augmentation. The planimetric changes are also not uniformly distributed over the entire study length. River width is the only feature that exhibits a significant difference between the pre-impact and post-impact periods with a clear progressive increasing trend, particularly at specific channel sections. While braiding and sinuosity also exhibit temporal changes, there is still much fluctuation from one year to another. Nevertheless, there has been a general increase in plan-form complexity of the river from a pre-dominantly sinuous form to a more braided and multi- channel pattern in the past 41 years. Avulsions resulting in elimination and straightening of the channel at many places along its length was a common phenomenon witnessed in the riverine system. Although the overall time-scale of analysis from 1973 to 2014 is small in comparison to channel change studies carried out across the world for other alluvial rivers, yet the degree of variations or adjustments shown by Dikrong within a short span of time is reasonably extensive.

Frequent changes in river morphology exacerbated by development projects such as dams and reservoirs pose both environmental problems on the riparian floodplain as well as socio-economic impacts on the riparian settlements adjacent to the bank. The consequences are more serious in rivers that suffer from the twin impacts of flow obstruction (by damming) and diversion (by water transfer to an off-site powerhouse) for electricity generation. In the next chapter, I explore these socio-economic aspects of river damming and flow deprivation upon downstream riparian communities in the Ranganadi basin.

103

Chapter 4

Socio-economic impact of the Ranganadi hydel project on downstream riparian populations in the Ranganadi floodplain

4.1 Introduction

Loss of forest, agricultural and settled lands due to reservoir submergence along with involuntary displacement, resettlement, and rehabilitation of upstream riparian populations are the common socio-economic consequences of hydropower development in any river basin (Iyer, 2003; Saikia, 2012). A vast majority of discussions on the socio- economic impacts however concentrates on people living upstream of the dams, whereas downstream populations are also known to suffer significantly. The nature, extent, and consequences of such downstream socio-economic impacts largely depend upon the respective communities’ dependence on the riparian ecosystem goods and services such as river-based livelihoods (Richter et al., 2010; Vagholikar and Das, 2010). For instance, the modification of flood regimes – magnitude, frequency, and timing, downstream from the dam can have serious implications on the adaptive capability of traditional cropping systems and other associated flood-plain livelihoods to natural flow patterns (Adams, 1985; Olukanni and Salami, 2012).

Thus, this chapter discusses the fourth category of impacts beyond the ecological impacts of dams, i.e., the impacts upon the social and economic sustainability of the riparian populations living alongside the impounded river - Ranganadi, in the downstream areas. It continues from the physical and hydrological changes that have occurred in the downstream riverine system detailed in Chapters 2, and discusses how (or whether) those changes have affected the riparian communities residing in the Ranganadi floodplain zone of Assam. The impacts discussed herein are exclusively ‘off-site impacts’ since the study is situated in the floodplain region approximately 30 km away from the site of intervention and not in the immediate vicinity of the dam up in the mountainous terrain of Arunachal Pradesh. In this chapter, I also discuss the different perceptions of the riparian communities towards the hydel project and the changes in the riparian ecosystem post- dam construction as observed by them. Since river dependence influences the probability and impact of riverine changes and hazards upon a particular riparian population, this study also examines the use and dependence of the sample riparian communities upon the

105

Chapter 4

ecosystem good and services, their livelihood structure and how (or whether) these have been impacted by dam-related changes in the river.

4.2 Data and methodology

Both primary and secondary data were used to examine the socio-economic impacts on the downstream riparian communities along Ranganadi. Primary data were collected through intensive semi-structured and open-ended questionnaire based interviews in selected downstream floodplain villages. The questionnaire addressed the riparian communities’ river dependence and livelihood practices with special reference to river- based livelihoods in the downstream floodplain areas and their opinions regarding the upstream dam.

The six sample villages of Ranganadi Basin that were selected to study the dam- related socio-economic impacts are namely –Borbil, No.1 Pachnoi Ujani and Ujani Khamtibaligaon on the left bank; and No. 2 Dejoo Pathar, Ujani Khamti and Pahumora on the right bank. All the villages belong to the Lakhimpur District of Assam. A total of 311 households were surveyed across all six villages. Interviews were conducted with the head of the household and in cases where the head was absent, the next person present who was capable of giving household details and answer the questions knowledgeably were interviewed. Analysis and interpretations provided on the changes in agricultural conditions, land ownership, fishing and other forms of riverine dependence are provided in terms of the household. However, interpretations on the perception towards the upstream dam are made in terms of the interviewed person and not for the household as a whole.

Data was also gathered with respect to the notable biophysical and ecological changes in the riverine ecosystem post-dam construction as witnessed by the floodplain communities. The extensive empirical data thus collected were subjected to both quantitative and qualitative analyses to ascertain the perceived and observed impacts of the Ranganadi dam on the downstream floodplain population. The economic and social changes between the pre-dam and post-dam periods were evaluated using descriptive statistics such as frequency and percentage analyses in IBM SPSS Statistics 23. The economic changes and the impact of the dam have been discussed as changes observed in the ecological goods and services provided by the river and the riparian ecosystem, post-

106

Chapter 4

dam construction, such as – effects on floodplain agriculture, fisheries, livestock, driftwood etc. Focus has been placed primarily on the changes in the livelihood structure of the riparian communities resulting from the physical changes in the river and alterations in downstream river flow. Within the social impacts, this study mainly discusses the changes in river dependence of the floodplain communities, migration due to the environmental changes and impact on livelihoods etc.

Based on the nearness to the dam in terms of the longitudinal distance, the villages on the two banks have been arranged and coded as given in Table 4.1. Fieldwork was conducted in two phases – the first during March to August 2014 and a second phase from September to December, 2014. The geographical extent of study within each basin was restricted to 6 km across the river (lateral zone of study) with 3 km on each bank and downstream extent (floodplain zone) as 20 km. The lateral distance of the studied villages pertaining to both basins in concern varied between 0.1 km and 3 km. While the longitudinal impacts of dam construction on a river can be quite extensive and continue for hundreds of kilometers downstream (Braatne et al., 2008; Grill et al., 2015), the lateral extent of impacts can remain restricted to a much shorter distance on both sides of the river. In their study, ‘Lost in Development’s Shadow: The Downstream Consequences of Dams’, Richter et al. (2010) narrowed down the lateral extent and subsequently the number of riparian people to be affected to a 10 km radius around the river. Their assumption was that people living within a 10 km radius are more dependent upon the riverine ecosystem for their livelihood activities and hence are at a higher risk. Using a similar assumption the lateral extent of the sample village selection was restricted to 3 km.

The major obstacle faced in comparing the social and economic changes was the lack of baseline information prior to dam construction. An attempt was made to study the impacts of the project as a temporal change in livelihood strategies by separating the livelihood activities of a particular household between pre-dam and post-dam time- frames. The questions were designed to collect information on livelihood activities both prior to 2002 and at the current time. The dam’s impact was analyzed as any change in livelihood activity that has occurred post-2002 and if the reason cited for the change was related to the dam. However, the data have some drawbacks as the pre-2002 information given by the respondents are based on memory, i.e. recall data. Nevertheless, in the

107

Chapter 4 absence of baseline or past data on livelihood activities of the surveyed households, memory recall is the only way in which changes and impacts in livelihood strategies could be ascertained. Similar use of recall data to determine pre-dam conditions in the absence of a baseline study was also employed by Adams (1985) in the study, ‘The Downstream Impacts of Dam Construction: A Case Study from Nigeria’.

Table 4.1: Codes assigned to the downstream villages that were surveyed along the two banks of the Ranganadi and the number of households interviewed in each village

No. of No. of Left Bank Right Bank Codes surveyed Codes surveyed villages villages households households

No. 2 Dejoo Borbil RL1 100 RR4 41 Pathar

No. 1 Pachnoi RL2 45 Ujani Khamti RR5 40 Ujani

Ujani RL3 35 Pahumora RR6 50 Khamtibaligaon

4.3 Background of the study villages

According to the various agro-climatic zones, the Ranganadi floodplains fall within the North Bank Plains Zone (NBPZ) of the Brahmaputra Valley. The fertile alluvial soils facilitate extensive cultivation of rice, rapeseed, mustard, wheat, jute, potato, and other vegetables. Likewise, net sown area covers more than 60% of the total geographical area in all the surveyed villages of Ranganadi Basin. The population is primarily agrarian and quite disadvantaged economically. Rice/Paddy is the dominant crop cultivated followed by potato, mustard seed and vegetables (rabi foodgrains). Paddy cultivation includes three varieties - Autumn (Ahu) rice, Winter (Sali) rice and Summer (Boro) rice. Winter rice is the most commonly cultivated variety across Assam and more so in the floodplains of NBPZ due to the abundance of rainwater (Goyari, 2005).

108

Chapter 4

Thus, in all the study villages, winter rice with sowing time in June-July and harvesting in October-November (duration 130 to 160 days average) was found to be the predominant form of paddy cultivation. Sali rice, typically grown in low-lying areas, can withstand water depths up to 30 cm. Amongst the different types of Sali cropping system, deep water or floating rice called Bao can endure water depths greater than 100 cm (Ahmed et al., 2011). Autumn and summer paddy was found to be less prominent. Besides paddy, rabi cropping can be witnessed during the lean rainfall months of October to March that involves cultivation of potato, vegetables, rape seeds, mustard etc. Other agriculture allied livelihood activities include dairy, goatery, duckery, poultry, piggery, and fishery. Private ponds and fisheries serving both commercial and self-consumptive purposes are also a common feature observed in the surveyed villages.

Despite being a primary livelihood practice, agriculture is still largely practiced in traditional forms, lacking modern technological inputs and with high vulnerability to climatic changes and flood damages. None of the sample villages has any irrigation facility and agriculture is primarily rain-fed (Census of India, 2011). Power supply is available only for domestic purposes and none for agricultural or commercial purpose. The region is greatly affected by the problem of recurrent floods due to an ‘abundance of water, particularly in the monsoon season’ with maximum impact upon paddy cultivation (Sharma, 2006; Northwest Hydraulics Consultant, 2006; Neog et al., 2016). ‘Multiple waves of floods’ and especially flash floods during the monsoon cropping season not only affect the standing crop but also result in permanent loss of productive agricultural lands through excessive silt and sand deposition (Neog et al., 2016). The left bank sample villages along Ranganadi in this study (RL1, RL2, and RL3) are examples of areas damaged by sand casting from overbank flows and embankment breaches in 2008 and 2009.

Rice is also worst hit by floods since these crops are mainly grown in low-lying areas that are inundated by flood waters. Additionally, the flood prone period of May to October coincides with the time of harvesting of autumn ‘ahu’ rice and cultivation of the winter ‘sali’ rice (the principal kharif crop), increasing crop vulnerability and damage (Goyari, 2005). Farmers practicing rice mono-cropping in marginal or small farm holdings in Assam are therefore most affected by the annual floods and are economically poor. As long as adequate irrigation facilities are not developed in the region, people will

109

Chapter 4

continue to be dependent upon rain-fed agriculture. Lack of irrigation especially during the lean period also restricts the riparian communities from shifting their cropping patterns to flood-free seasons.

All the surveyed villages reported the presence of households leasing out and leasing in land for agricultural cultivation. Land is leased either for farming rice or rabi crops. There were households with marginal land holding size that leased in additional

land for increasing the existing area of paddy cultivation or cultivate rabi crops as a secondary crop variety along with paddy.

Table 4.2 details the demographic profile of the study villages. RL1, RL2, RL3, and RR5 had a sizable proportion of Scheduled Tribes population, with majority of them belonging to the Mishing community. The Mishing community is one of the major ethnic groups in the north-bank Brahmaputra plains. The community has long been known for their close association with rivers and their livelihood dependence upon the riparian ecosystem19. Historically too they are commonly known as the ‘river people’. The name – Mishing itself denotes people living by the river banks as ‘Mi’ means ‘man’ and Shing means ‘water’ or ‘river’, thus establishing the strong traditional bond between the community and the riverine ecosystem (Das et al., 2009).

Another distinct feature of the Mishing community is the structure of their houses that have been indigenously designed to withstand floods and earthquakes. These traditional houses known as ‘Chang ghar’ are elevated about 6 ft. above the ground on bamboo or wooden stilts, with thatched or tinned roofs and bamboo or wooden floors (Fig. 4.1). The empty space underneath is used to shelter livestock (mainly pigs and poultry). Being well above the ground, these raised structures remain safe from submergence during floods. Additionally, the flight of stairs leading up to the raised platform has much religious, cultural, and social significance attached to them and any guest or bride is truly accepted into the family only when led up these stairs (Rural Volunteers Centre, n.d.). Similar to the raised dwelling structure, the granaries are also built on stilts and elevated above the ground for protection from floods and wild animals. Raised granaries are a traditional practice of other non-Mishing Assamese communities

19 The dependence of the Mishing community upon the riparian ecosystem and their traditionally developed adaptive capability to riverine hazards are discussed in detail later in ‘Section 4.6.3 Changes in riverine use and dependence’.

110

Chapter 4 as well and could be seen throughout the villages. Apart from the Mishings and the general caste population, other communities living in the study villages include the Adivasis and the Ahoms.

Table 4.2: Salient demographic features of the six surveyed villages located along the Ranganadi. The distance of the village from the river and the dam has been approximated from GPS points collected during the field survey and Google Earth. The demographic information is as per the Census of India (2011) data (District Census Handbook, Lakhimpur). Borbil No. 1 Ujani No. 2 Ujani Pahumora (RL1) Pachnoi Khamti- Dejoo Khamti (RR6) Village Ujani baligaon Pathar (RR5) (RL2) (RL3) (RR4) River bank Left Left Left Right Right Right Lateral distance 1 – 2.5 0.5 – 2 0.2 – 1 0.5 – 1 0.1 - 0.5 0.5 – 1.5 from river (km) Downstream 38 45 50 42 46 52 distance from dam

(approximate) (km) Total geographical 315.33 214.74 312.86 317.09 157.72 290.68 area (in ha.) Total no. of 362 137 105 251 146 195 households (HH) Total 2056 693 504 1449 704 950 Population Male 1083 365 261 750 367 479 Female 973 328 243 699 337 471 Total Scheduled Castes (SC) 0 4 5 0 0 80 population Total Scheduled Tribes (ST) 866 335 199 0 516 5 population Total 1141 463 391 778 449 733 persons Literates (nos.) Male 635 266 211 441 255 395 Female 506 197 180 337 194 338 Total Workers Main 681 153 43 270 160 230 (nos.) Marginal 125 205 131 154 64 21 Main 473 141 8 139 117 168 Cultivators (nos.) Marginal 24 94 82 56 35 5 Agricultural Main 115 8 0 53 16 1 labourers (nos.) Marginal 52 111 9 35 13 6 Household Main 9 0 0 13 0 1 industry workers (nos.) Marginal 17 0 4 11 0 0 Other workers Main 84 4 35 65 27 60 (nos.) Marginal 32 0 36 52 16 10

111

Chapter 4

Fig. 4.1 A typical chang ghar as seen in Borbil

4.4 Key geographical features of the study villages

Topographical features and the location of a particular floodplain village with respect to the river play an important role in the spatial distribution of river-related hazards. While all the six study villages are similar to each other in the overall land-use and land-cover conditions, yet certain topographical features such as embankments and roads, sub- channels of the main Ranganadi etc. differentially distribute the impact of riverine hazards across the villages as well as within the same village. An important topographical feature of RL2 is its location as a vegetated and built-up riverine island between the main Ranganadi channel and a sub-channel of the river locally named – Joyhing huti (Fig. 4.2A). The Joyhing sub-channel is formed upstream of RL2 at a point near RL1 where it initially branches off the main channel and rejoins it approximately 8 km downstream near RL3. The modification of the Joyhing over the years and especially after the flash floods in 2008 has been explained in detail in Chapter 220.

Originally, this sub-channel cut across the riparian landscape as a small stream that was an important source of water and fish for the village. Significant hydrological and physical modifications occurred in the sub-channel following the 2008 flash floods whereby the channel had become wider and shallower due to heavy sand deposition. In 2008 when flood waters were suddenly released from the upstream Ranganadi dam, a considerable amount of the flow also shifted towards the sub-channel and given its

20 See section ‘2.4.4 Capture of the Joyhing channel and diversion of flow in the post-dam period’, Chapter 2. 112

Chapter 4

narrow size, bank overflows resulted in heavy flooding and sand casting21 in the adjoining lands. As a result, the sub-channel became wider and shallower (given bed aggradation due to heavy silting) and in the subsequent years, it transported significant amount of water and sediment along with the main channel causing floods in newer plains and

Fig. 4.2 Location of RL1 (Borbil) and RL2 (Pachnoi Ujani No.1) villages along the left bank of Ranganadi, (A) inset showing how RL2 is located between the Ranganadi and the Joyhing in the form of an island bar and thus is affected by flow fluctuations in both channels. The image date is March 7, 2018. (B) A second inset of RL2 dated July 27, 2006 is prior to the flash flood in 2008. It can be observed that the Joyhing although present is much smaller compared to the one in the 2018 image. The open green patches in the images represent agricultural lands (primarily under paddy) (Source: Google Earth Pro).

21 Sand casting is the ‘deposition of large amounts of sand by flood water’ (Das et al., 2009). 113

Chapter 4 agricultural lands that were earlier relatively flood free. At the same time, it remains completely dry and sandy during the lean flow season owing to obstruction of upstream flow by the RHEP dam. Locals in RL1 claim that the drastic fluctuation in flow between the wet and dry season months has significantly reduced fish populations and water availability in the sub-channel and adjoining wetlands.

Similarly, in RL1, part of the village comprising some 50-60 HHs is situated in the continuous floodplain between Ranganadi and the Joyhing sub-channel unobstructed by embankments, while the rest of the village is situated away from both channels (on their east) and protected by embankments. Although the embankment acts as an obstruction to river related hazards, yet on multiple occasions they have caused more damage than benefit due to breaches such as in 2008 and 2009. The fertile floodplain between the main channel and the sub-channel had come under direct impact of the 2008 flash floods and the resultant sand casting completely changed the riparian landscape, the impacts of which still persist.

All three right bank villages (RR6, RR5, and RR4) are situated between 200 meters to 2.5 km from the river. A distinct feature of the right bank is the presence of the highway leading to Arunachal Pradesh, which fragments the floodplain into two parts. In addition, the road leads upstream to Yazali where the Ranganadi hydroelectric project (dam and reservoir) is situated. As the highway acts as a barrier to the direct onslaught of river-related hazards, there is marked difference in the impacts of riverine changes on people living on either side of the road. Floods and other such changes in the riparian ecosystem more directly affect people living on the east of the road and by the side of the river, while people living on the western side of the road are less affected. Yet, the surveyed households in RR6 did report that when the water level in the river increases beyond the danger limit, some of it gets pushed back upstream on its right bank through various points of entry and breaks in the embankment protection. Usually the agricultural fields in a floodplain are connected to each other and wherever there is a road separating the fields, water flow between the fragmented lands is facilitated by installing bridges or drains. It is through such entry points in case of the Ranganadi’s right bank that flood waters from the river reach the agricultural lands and settlements otherwise protected by the highway, thus affecting them but to a lesser extent. The significance of topography as

114

Chapter 4

a controlling variable of impacts from riverine changes is discussed further under section 4.6.5 (b).

4.5 The Impact of June 2008 flash flood

The Ranganadi hydel project is a single-purpose project built solely for electricity generation without flood control or irrigation objectives. In a region where floods are a persistent problem and an annual occurrence given heavy monsoon precipitation during the months of June to September, construction of hydropower projects with no flood control measure is a major lacuna in sustainable project planning. The 405 MW Ranganadi hydel project has a concrete gravity dam of 68 m height and 344.75 m length (at the top), thus attributing it with a relatively ‘small pondage’ (Ministry of Power, 2008). At the same time, 160 m3/s of water is also diverted to the adjoining Dikrong River for electricity generation, which gives rise to a whole set of other problems in itself that have been discussed in Chapter 3 and 5.

On June 14, 2008, owing to cloudbursts and heavy rainfall in the upper catchment areas of the Ranganadi, there was a large volume of inflow into the reservoir and associated floods downstream with multiple peaks. Instead of a slow release of the flooding waters into the downstream channel, the water was held back for as long as possible, thereby attenuating downstream flow, but in the process increasing the risk of dam overflows. The excess water was finally released to the downstream channel without any warning to the floodplain populace after a certain point when the retained waters threatened to jeopardize the dam structure. Ultimately, the sudden and unanticipated release resulted in devastating flash floods in the downstream riparian villages, with high loss of property and crops. Besides the immediate damages to property, livestock, and standing crop, the floodwaters had also resulted in widespread sand casting over fertile agricultural lands whose adverse impacts continue to be witnessed till date.

The 2008 floods had more impact upon the left bank study villages along the Ranganadi and were crucial in influencing the perception of the riparian communities towards the upstream hydel project, as they held the dam responsible for their losses. Most of the physical changes with respect to the left bank riparian landscape and hence economic changes in the studied villages was reported to have occurred post-2008 rather than in the immediate years post dam construction. One of the highly affected left bank

115

Chapter 4 villages was Borbil (RL1), where the loss of property (dwelling houses, granaries, orchards) and agricultural lands resulted in the internal displacement of a number of families. Most of the displaced families were living outside embankment protection on the stretch of land enclosed in between the main Ranganadi channel and the smaller Joyhing side-channel. Some of the families whose houses were washed away by the flash flood shifted to other parts within the same village, especially near the village primary school right beside the embankment (Fig. 4.3A and B). Having lost their own house and land to flood and sand deposition, the families are now residing on government land and are afraid of going back fearing the repetition of similar flash floods like the one in 2008.

Fig. 4.3 Images showing the long-term socio-economic effects of the flash flood in 2008 in RL2 village, (A) temporary settlements along the embankment, of some of the families that had been internally displaced during the flood, (B) part of the primary school in the village that was damaged by flooding waters that had breached the embankment in 2008 and 2009, (C) and (D) past and present living conditions of one of the affected and internally displaced families. The previous house was built on concrete stilts unlike the present one and came to be completely covered by coarse sand deposition in the aftermath of the flood.

The primary school itself sustained much damage to its structure when the flooding waters that had breached the nearby embankment broke parts of the school (Fig. 4.3B). Due to lack of immediate and proper repairing of the breach, further damage was caused by the subsequent floods of 2009. Figure 4.3C and D shows the living conditions of one of the affected villagers in RL1 before and after the floods.

116

Chapter 4

In addition to the left bank villages, the 2008 floods had also impacted upon a few households in RR5. About 43% of the total surveyed HHs in RR5 had farmlands also located across the river in the left bank villages of RL2, RL3, and another village - Kalabeel. Out of these households, 35% reported being affected by the flash flood and ensuing sand casting over agricultural lands on the opposite side of the river.

4.6 Results and discussion

4.6.1 Impact on agricultural lands and the resultant changes in cultivation

Agriculture being the major economic activity in the surveyed villages, net sown area constituted the maximum land-use and land cover in the riparian floodplain. According to Census 2011, net sown area covers more than 90% of the land area in the villages of RL2, RL3, RR6, and RR4. However, various adverse impacts on the quantity and quality of agricultural lands were reported by the surveyed population from impacts of both natural climatic and dam-induced flow changes in the river.

The major change noted in terms of agriculture between the pre-dam and post- dam periods was the increase in the proportion of households leasing in land for agriculture post-dam construction especially in RL1 and RL2. Figure 4.4 displays the percentage distribution of households in each village leasing in agricultural land during the pre-dam and post-dam periods. RL1 exhibited the highest change from 16% to 54% post-dam, while RL2 exhibited an increase from 48% to 60% in the post-dam period. While the observed increase could be attributed to changes brought by natural processes of agricultural extensification to increase production (a case more prevalent in the right bank villages), the increases in the left bank villages of RL1, RL2, and RL3 were specifically attributed to the impacts of the upstream dam. The impact of persistent floods (both natural and dam-induced), sand casting, and embankment breaches on agriculture has been discussed repeatedly. Likewise, sand casting was reported as the chief cause of partial and at times complete loss of land by majority of the households in RL1 (76%) and RL3 (70%) (Fig. 4.4). In fact, extensification occurred primarily after the devastation of 2008 flash floods in case of RL1 whereby out of the 54% households, 39% began leasing in land for paddy or rabi cropping post-2008.

117

Chapter 4

% of households of %

Village Fig. 4.4 Percentage distribution of households leasing in land for agriculture as well as affected by sand casting in each of the surveyed villages during pre-dam and post- dam periods.

As mentioned in section 2.2, a part of RL1 is located between the main Ranganadi and the Joyhing sub-channel. This village niche was most affected in 2008 with significant changes to the overall landscape. Besides loss of standing crops, the floodwaters deposited at least 3 to 4 feet of sand upon the agricultural lands. Sand affected lands take at least 8 to 10 years to regain productivity and be suitable again for cropping, especially paddy (both summer and winter rice), unless the sand can be excavated out at own cost. Production still remains low for some time. Even for parts of the village that were likely to be protected by the presence of the embankment, breaches at two points affected most agricultural and settled areas of the village. Nevertheless, one of the short-term alternative income sources that resulted from the large-scale sand deposition was the selling of the sand. Still, it is profitable for only those lands that were deposited by a sand layer of at least 3 to 4 feet since such sand is suitable enough to be used for construction purposes and can be excavated with less difficulty. Greater losses were incurred in cases where neither the level of sand deposition was high enough to be profitably sold, nor low enough to be fit for cultivation.

In RL2, multiple households reported irregular flooding which they perceived to have increased in frequency and severity post-dam construction as the major cause to crop damage. Hence, most of them started leasing land for cultivation in areas that are lesser flood prone.

118

Chapter 4

The problem of sand deposition over agricultural lands was, however, observed to be quite low in the right bank villages of RR4 and RR6. The latter was also the only surveyed village that exhibited a decrease in households leasing land during the post-dam period. This right bank village of Ranganadi displayed a comparatively lower proportion of primarily farming based households and an alternately high share of non-farming HHs (Table 4.3) during both time-periods. At the same time, RR6 exhibited comparatively much higher presence of livelihood diversification with 68% of the HHs already practicing multiple livelihood activities before dam construction (Fig. 4.5). Sand casting impact was reported by a notable 43% in RR5 mainly due to the 2002 and 2008 flood occurrences. This right bank village had suffered flooding in 2002 due to a breach in the right embankment of Ranganadi (Fig. 4.6).

households of %

Village Fig. 4.5 Village-wise percentage distribution of households having multiple livelihoods during pre-dam and post-dam periods.

A significant area of low-lying paddy farmlands located by the village and the embankment was submerged for a prolonged period following the breach. There was no outlet for the water to return to the river and later on, the affected area got transformed into a swamp unsuitable for cultivation. The swamp still existed at the time of field survey in 2014. Therefore, 25% of the households in RR5 cited 2002 as the year of impact. However, the 2002 breach and flooding were not dam-induced as could be observed from Fig. 4.7, comparing the hydrographs of 2002 and 2008 (May to September daily flow). Firstly, the 2002 peak (401.5 m3/s) is much lower than the peak in 2008 (695 m3/s) and secondly, there was a gradual increase and decrease in daily flows preceding

119

Chapter 4 and following it. A completely different trend is visible in the 2008 flood peak, which is sudden with an abrupt rise and fall. Likewise, the 2002 occurrence was possibly not flash flood related, but more the result of a weak embankment structure that was breached at the time the river was flowing at bankfull stage.

Fig. 4.6 Embankment breach on the right bank of Ranganadi in 2002 and the ensuing inundation of agricultural lands in RR5.

Fig. 4.7 Hydrographs of daily flow for the years 2002 and 2008, pertaining to the high rainfall period of May to September. Both years were reported as impact years by the surveyed population in RR5 when long-term changes to agricultural lands had occurred, triggering changes in livelihood.

120

Chapter 4

4.6.2 Impact on the livelihood structure of the downstream communities

4.6.2.1 Changes in primary livelihood activity

Besides increased leasing of land for cultivation, the impacts of dam-induced flash floods and sand casting also resulted in changes in the livelihoods of downstream communities. In all the villages as represented in Table 4.3, agriculture and agriculture-allied activities, i.e., farming constituted the primary livelihood source in majority of the households during both the pre-dam and post-dam periods. Although all the six villages except RR5 exhibited a decrease in the percentage of households having farming as a primary livelihood activity in the post-dam period, it was not a notably high change. RR5 did not exhibit any change in primary livelihood between the pre-dam and post-dam periods despite being located less than 0.5 km from the river. The post-dam decrease in primary farming households across the sample villages was accompanied by a simultaneous increase in the percentage of households pursuing non-agricultural activities and wage labor as a primary livelihood source. This change could be attributed to impact of the upstream dam upon the agricultural fields and crop productivity as well as to changes brought by other normal processes. However, farming still continued as a secondary livelihood activity instead of complete ‘farm-exit’, even for households that expanded their livelihoods to other non-agricultural activities.

Table 4.3: Village-wise distribution of households (HHs) according to their primary livelihood activity between pre-dam and post-dam periods (in percentage). The primary livelihood activity has been broadly distinguished into three categories based on whether a particular household is engaged primarily in floodplain agricultural activities, or other non-farm and wage/casual labor livelihoods.

Village Pre – dam primary livelihood Post – dam primary livelihood activity activity Farming Non- Wage Farming Non- Wage (%) farming labor (%) (%) farming labor (%) (%) (%) RL1 93 3 4 87 5 8 RL2 97 0 3 92 3 5 RL3 82 6 12 67 18 15 RR4 44 39 17 27 49 22 RR5 75 10 15 75 10 15 RR6 48 44 8 44 48 8

121

Chapter 4

RR4 exhibited the highest decrease in primary farming households from 44% (pre-dam) to 27% (post-dam), yet except for a couple of households none of the shift in post-dam primary livelihood could be attributed to any impact of the dam on the riparian ecosystem. This is also supported by the fact that the overall impact of dam-associated river flow changes upon agriculture based livelihoods as perceived by the residents of RR4 is low. This right bank village is not as affected by problems such as sand casting as the other villages, whereby only 5% of the households attributed the changes in their livelihood structure to such river-induced hazards. Although flooding has been intensified by sudden and irregular releases from the dam, yet the problem of floods had always existed in the area, hence, people are accustomed to the exposure and livelihoods have not changed as such. The surveyed riparian communities of RR6 reported similar coping mechanisms, where they exhibited a resilient and adaptive capacity to floods owing to long-term past exposure to such occurrences. In RR6, 72% of the surveyed households asserted that flooding has been a long existing problem in the area even prior to dam construction. However, 54% of the households did also claim that sudden, irregular, and untimely releases of stored reservoir water by the dam have exacerbated the flooding problem. At the same time, the RHEP dam lacks flood control mechanisms in a highly flood prone region with heavy monsoon flows and the possibility of re-occurrence of events such as the flash flood of 2008 still exists. Under such circumstances, the resilience and adaptive capability of villages like RR6 against unexpected flow changes with long-term impacts is unknown.

Thus, it was observed that although agriculture is significantly affected by problems of flood and sand deposition, yet there has not been any notable change in the primary livelihood structure. Instead of a ‘farm exit’ (Bhandari, 2013) phenomenon, adaptation strategies such as livelihood diversification, agricultural land extension by leasing additional land, shift from paddy to rabi cropping were more prevalent. Agriculture as a primary livelihood source has been a long-established tradition in the small and marginal farming households of the downstream villages of the Ranganadi Basin. Only extreme changes such as complete loss of land would usually lead to an abandonment of traditional practices (example RL1), which could be a possible reason for the sustenance of farming despite the degradation of the river and its riparian ecosystem. The indigenous and traditional cropping systems in the floodplain has evolved according to the distinct precipitation patterns of region and resonate with the seasonal flooding and

122

Chapter 4

drying of the rivers. However, anthropogenic interventions such as hydropower projects threaten the sustainability of such traditional methods of cultivation by altering the natural flow regimes of the rivers and disrupting the natural flood cycles to which the cropping systems have long been attuned.

On the other hand, in the right bank villages wherever livelihood transition was reported, it was not necessarily attributed to the dam. The transition resulted more due to other social, economic, and familial factors unrelated to the presence of the dam or its effects on the riparian ecosystem.

In general, livelihood diversification is the more prevalent risk management strategy and coping mechanism against economic and environmental stress observed in rural households engaged in subsistence farming or small farm wage labor (Ellis, 1998; Bhandari, 2013; Gautam and Anderson, 2016). Similarly, in the studied villages, addition of or diversification to other non-farm livelihoods while still practicing farming as a primary or secondary livelihood activity was the more prominent response witnessed to post-dam impacts upon agriculture thereby increasing security, resilience and adaptive capacity of a particular household. While multiple livelihoods existed notably even prior to dam construction, yet there was much increase in the percentage of households with multiple livelihoods in the post-dam period in RL1, RL2, RR5, and RR4.

4.6.2.2 Livelihood diversification and adoption of multiple livelihoods

Figure 4.5 shows the percentage of households engaged in more than one livelihood before and after dam construction. Most of the households in all the 6 sample villages were involved in multiple livelihood activities even prior to dam construction. However, there was an increase in adoption of multiple livelihood strategies post-dam commissioning. This increase was observed to be highest in RL1 and RL2 whereby percentage of households having multiple livelihood activities had increased from 32% to 54% in the former and from 33% to 60% in the latter. Farming formed an association with other livelihood activities as both a primary as well as secondary source of livelihood.

The higher increase in multiple livelihoods in RL1 and RL2 as compared to other riparian villages of Ranganadi could be attributed to the adverse effects of sand deposition on agricultural lands (changing both quality and quantity of land) in 2008 and

123

Chapter 4

2009. Majority of the households in both villages were observed to have switched to multiple livelihoods to compensate the loss in paddy farming. There were also households that gave up farming altogether due to damages to either owned or leased in land post- 2008. In a few cases, cropping was attempted on the affected lands, but the yields were too low to continue farming in the subsequent years. In RL2, the change was also prompted from adverse and irregular flooding of the rice fields and resultant economic losses from crop damages.

There was negligible change in the percentage of households adopting multiple livelihood activities post-dam construction in RR6. Before the dam had come up, 68% of the households already had more than one livelihood source, which changed to 70% post dam.

Besides livelihood diversification, within farming households an adaptation strategy that was observed was the shift from paddy to rabi cropping, or the addition of rabi cropping along with paddy during the lean flood-risk free season in the post-dam period. A predominant practice in small and marginal farming households of the NBPZ is single-season mono cropping of rice (Neog et al., 2016). Winter rice being the dominant crop typically occupies farmlands from June to December. Once it is harvested, most of the farmers keep their fields fallow, instead of sowing a second crop. The rest of the non- paddy season is mostly dedicated to other forms of livelihood activities such as wage or casual labor or other miscellaneous jobs. Double cropping, which entails cultivation of a second crop on the same plot of land where rice is cultivated, is a beneficial practice that increases income of the farmers as well as enriches the soil with nutrients for the next phase of rice cropping. Yet, majority of the small and marginal farmers across Assam do not practice double cropping. Recently, in an attempt to encourage double-cropping (with pulses and oilseeds) which is more profitable and increases soil-fertility, the Agricultural Department of Assam, has started a centrally funded program named Targeting Rice Fallow Area (TRFA) (in eastern India), to cover rice fallow areas of over 15,000 ha in three of the state’s districts - Nagaon, Golaghat, and Sivasagar (Thakur, 2017) 22. Such programs need to be extended to other districts like Lakhimpur, where rice cultivation is largely affected by floods and double cropping would increase the financial resilience of the otherwise affected farmers.

22 For details see ‘3 state districts brought under double cropping’, The Assam Tribune, 25 February 2017. http://www.assamtribune.com/scripts/mdetails.asp?id=Feb2517/at056. 124

Chapter 4

4.6.2.3 Impact on downstream fish availability and on the practice of fishing

In all the sample villages, fishing was claimed to be practiced only as a tradition and seldom on a commercial scale or as a primary source of livelihood. Yet in whatever form it might be, the practice still establishes the close association of the riparian communities with the floodplain ecosystem and their river-based dependence. Besides the Ranganadi and its sub-channels, the adjoining wetlands or ‘beels’, natural or artificial fishponds and tanks, swamps and other low-lying areas also served as active fishing grounds for the riparian communities. Wetlands, typical to the Brahmaputra floodplains, have high ecological significance as they support the habitats of a wide variety of flora and fauna. Through an exchange of water and sediment with the river, these inter-connected systems work as ‘flood retention basins and traditional fisheries’ (Goswami and Das, 2003). Wetlands are also significant from a cultural perspective since they act as points of community fishing actively participated by women. Similarly, individual and community fishponds and tanks are a common feature in the floodplains of Assam and in Lakhimpur district alone there are about 10,774 nos. of such ponds covering a water spread of 1996 ha. (Government of Assam, 2015). In case of RL1, 15% of the total surveyed households owned their own fishery or pond whereby spawning and rearing of the popular fish varieties constituted both self-consumptive and commercial purposes. Therefore, apart from the river multiple other sources for fish could be observed in the surveyed villages. However, the wetlands and other low-lying areas in the floodplain form a closely inter- linked system with the river and are greatly influenced by the hydrological patterns of the riverine system. Even, individual fisheries situated nearby to the river came under the influence of riverine changes whereby during floods the overbank flows tend to wash away the reared fish or result in sand casting in the fishponds.

Accordingly, fishing was found to be an active practice in RL2, RL1, and RR5 with respectively 70%, 72%, and 80% of the surveyed households reported traditional fishing activities at some point in their lives (Table 4.4). RR6 reported the lowest proportion of households practicing fishing with many exhibiting reluctance in accessing the main river channel (on the opposite side of the highway to Arunachal Pradesh) and depended more on procuring the protein from available local markets or from a nearby smaller stream - Singora. Nevertheless, there was a change in the practice post-dam construction as multiple households reported abandoning fishing due to various reasons.

125

Chapter 4

Table 4.4: Village-wise distribution of households (in percentage) who have practiced fishing in the riparian ecosystem either prior to or post-dam construction and the changes in riverine and riparian fish availability post-dam according to their observations (Changes in fish availability have been opinionated by HHs irrespective of whether they practice fishing or not).

HHs reporting… HHs who have No change in fish practiced fishing Post-dam decline quantity between Village ‘Don’t know’ either pre- or in fish availability pre-dam and (%) post-dam (%) (%) post-dam periods (%) RL1 72 76 3 21 RL2 70 68 0 32 RL3 51 73 0 27 RR4 66 63 5 32 RR5 80 75 2 23 RR6 36 38 4 58

Based on the repetition of major reasons and observations cited by the surveyed households that were noted to be analogous across the sample villages, this study tries to establish the impact of the upstream dam upon downstream fish availability and further upon the riparian dependence.

Decline in both riverine and riparian fish populations was a common post-dam observation by majority of the surveyed households irrespective of whether or not they were engaged in fishing. Even though absolute numbers of the riverine and riparian fish population could not be collected from either primary or secondary sources, yet there is a consensus that fish populations have considerably decreased in the post-dam period. Rampini (2016) reported similar decline in fish abundance following construction of the hydro project indicating the adverse dam impacts. Some villagers in this study gave anecdotal evidence of the post-dam change in quantity of fish catches as -

“Earlier 3-4 kg could be caught from the river in an hour, while now only 250 g can be caught after an hour’s fishing. Fish populations in the river has decreased sharply post-dam” - villager in RR6, August 2014.

“1 to 1.5 kg could be easily caught from the river earlier, but now the quantity has decreased and fishing has become rarer” - villager in RL1, April 2014.

126

Chapter 4

“Plenty of fish was available earlier in the river. About 10 to 15 kg could be caught at one go using the ‘jaal’ (fishing net)” - villager in RL1, May 2014.

Table 4.4 presents the percentage of households that reported decrease in fish availability during the post-dam period as a consequence of dam-induced changes in river flow regimes. The primary reason cited for the decline in fish numbers across all the sample villages was the extreme decrease in river flow and drying up of the river channels especially during the lean period specifically post-dam construction. During winter, the river completely dries up leaving no space for the fish to survive, whereas when the dammed water is released, the velocity of flow becomes too high for the fish to survive or even be caught.

Irregular flooding and disruption in the exchange of water between the river and riparian floodplain has negatively influenced spawning and fish productivity, thus reducing overall fish availability. Majority of the population claimed reduction in wetland fishing following dam construction due to decreased fish productivity of these traditional fisheries that were drying up due to the decrease in river flow as well as excess sand deposition left behind by receding flood waters.

“Earlier (pre-dam) plenty of fish and turtles could be found in the adjoining wetlands and were enough for self-consumption (at least). However, these places gradually got sand deposited and the 2008 flash flood deposited around 7 - 8 ft. of sand” - villager in RL2, March 2014.

Another villager in RL2 claimed, “Earlier plenty of fishing was done in the wetlands. But ever since the river has been dammed, the wetlands have dried up and died (also sand deposited) and so no fish is available.”

A villager in RL3 observed that currently fishing in the Ranganadi and the adjoining wetlands are restricted to the rainy months of July/August, although availability has significantly decreased. A wetland known as ‘Kerahi beel’ situated between Joyhing and Ranganadi (close to RL3 near the current embankment position) was the common and a rich source of local fish. The mentioned wetland never dried up prior to damming of Ranganadi, even during the winter season, and fishing could be done the whole year round (even in the dry months of December). However, post-damming the wetland dried up.

127

Chapter 4

Dams disrupt the nutrient cycle of the aquatic ecosystem crucial for fisheries by altering the seasonal inundation of wetlands by the river (Menon et al., 2003). Goldsmith et al. (2003), discuss similar adverse effects of the Aswan dam upon downstream fisheries, whereby sardine catches had decreased by 97% post-dam construction due to reduced silt-loads and the resultant deprivation of essential nutrients. Shoemaker (1998) reported 30% - 90% decline in fish catches downstream of the Nam Theun-Hinboun dam (a run-of-river inter-basin water diversion project) in Laos (Mekong River Basin) due to drastic reductions in water levels of the impounded river. This ultimately had significant impacts upon the associated livelihoods.

The negative impact of dams on downstream fisheries is a widely documented phenomenon. However, bank protection structures such as embankments also result in significant reduction of fisheries as they cut off the feeding and spawning habitats of many of the fish species and also inhibit movement between the river and the floodplain (Baruah and Biswas, 2003; Menon et al., 2003). Similarly, in case of the Ranganadi, the RHEP is a major and definite causal factor of diminished fish availability. Yet, additional anthropogenic factors such as embankments and encroachment have also decreased downstream fisheries considerably. Reduction in fish catches were also reported by the surveyed households owing to degradation of wetlands through sand deposition and drought caused by reduced river flows. However, most of the wetlands also exhibit degradation due to human encroachment for settlement, pollution and eutrophication resulting from destructive land-use practices (Goswami and Das, 2003).

Dandekar (2012) asserts the urgent need for maintaining river-specific environmental flows and sediment flows in impounded rivers, installing fish ladders in hydro projects (large and small) and stronger policies and amendments to the existing Fisheries Law 1897/ Wildlife Protection Act/ Environment protection laws to ensure the same. The Indian rivers are endowed with a rich variety of fishes that maintain the ecological balance and provide crucial nutritional and livelihood security, especially to the rural poor. Yet, these rivers lack proper documentation of the status of riverine fisheries, population and status of dependent fisher folk, assessment of dam-induced changes on fisheries and the dependent fisher communities etc. (Dandekar, 2012). Despite a large population of the rural poor in India depending upon riverine fisheries, the focus of the National Fisheries Development Board is more on reservoir fisheries and

128

Chapter 4

aquaculture (Dandekar, 2012). Even in case of the Ranganadi and Dikrong Rivers, there was hardly any baseline information available on the status of riverine and wetland fisheries. Therefore, all inferences about the change in downstream fish numbers and fishing patterns post-dam obstruction of the Ranganadi were made based upon the information provided by the interviewed households in the villages.

4.6.2.4 Impact on animal husbandry including grazing patterns

Animal husbandry is the most common agriculture-allied activity prevalent in rural farming economy. Similarly, in all the surveyed villages, majority of the farming households and to some extent even the non-farming households reared livestock either for commercial or self-consumptive purposes. The most commonly found livestock were cows, goats, pigs, and poultry (chicken, ducks etc.).

General observation across all the riparian villages was the decrease in livestock rearing. Almost all the households reported a decreased number of livestock during the current time or less than what used to be the number a decade ago. Many accounted fewer numbers of individuals available in a family as the chief reason for difficulty in looking after the animals. As more members of a household shifted to other forms of non-farm livelihood activities, the number of livestock gets reduced. However, in villages such as RL1, the primary reasons cited were to reduce the risk of loss by reducing the livestock number given past flood experiences and the probability of re-occurrences of such events. Other reasons cited include – difficulty to look after the animals during the rainy season when monsoon floods arrive combined with scarcity of fodder and death of animals following consumption of silted grass. The lack of higher grounds and the difficulty in shifting the animals to such places where they could be guarded from the flooding waters also poses a hindrance to having a higher number of livestock. This phenomenon of disposing animals out of difficulty of carrying them to relief camps or dykes, instead of monetary gains, has also been reported by Mahanta and Das (2017) with respect to flood affected villages in Lakhimpur and Dhemaji districts, Assam.

However, the overall decline was not attributed entirely to the upstream dam and the dam-related changes in the river. Most people revealed that with increased population, more riparian areas, including the grazing areas and chapories were being encroached and converted to built-up land. There has also been a conversion of grazing lands for

129

Chapter 4

cultivation purposes, especially for rabi cultivation during the lean period when the decrease in river water levels clear up these areas further making them suitable for agriculture. This has led to a decrease in fodder and space for grazing the animals; whereby the livestock has to be reared on the farmlands when they remain fallow post- cropping.

Riparian residents in RL2 also claimed that the native grass commonly present in the riparian areas along the river have been replaced by certain invasive species that are not fed upon by the cattle. People believed that this to be a post-dam phenomenon and asserted that floods caused by dam releases deposit invasive plants on the downstream floodplain; plants that are carried down from the upper riparian. Similar observation was put forward by a resident in RR5 that the adjacent grasslands and char areas became vegetated by a plant variety (locally known as Birina), following the 2002 floods. This particular plant obstructs the growth of natural grass (used as fodder) underneath it causing an adverse ecological change. Whether the dam is actually responsible for such changes in riparian flora, however, has not been examined in this study as it remains beyond its scope.

4.6.3 Changes in riverine use and dependence

Development of the riparian lands and floodplains along a river for human settlement and agriculture has been going on since early ages. Floodplains, because of their fertile and nutrient-rich soil and assured supply of surface and ground water, form ideal grounds for fishing, settled agriculture and grazing. Besides these, the riparian communities also access the riverine ecosystem directly for drinking and domestic uses and collection of driftwood for fuel. Riverine use and dependence usually vary according to the proximity of a particular household to the river, whereby people living nearer to the river (even within the riparian ecosystem) are comparatively more dependent. On the contrary, further the distance from the river, lower is the dependence and use. At the same time, a particular household might stay some distance away from the river, yet would access the adjacent riparian areas more frequently and have greater dependence as they access the river banks and islands (chapories) for grazing livestock. Such households, as has been discussed, are also likely to be affected more by dam-induced changes in the riparian ecosystem particularly by floods and bank erosion. River dependence of the floodplain communities along Ranganadi was evaluated in terms of their use of the river for drinking

130

Chapter 4

and domestic water purposes, driftwood collection etc. between the pre- and post-dam time-periods in order to examine how it has evolved following dam construction.

Table 4.5 shows that use of the riverine water directly for drinking purposes was quite low even during the pre-dam period except for the people of RL1 (88% HHs) and RL2 (63% HHs). More than the surface water resource, ground water constituted the primary drinking water source for surveyed downstream villages on both banks of the Ranganadi. Open wells (concrete) and tube-wells were found to be a common source of both drinking and domestic water as well as water for livestock. Drinking water use decreased further after dam construction whereby in RL1 and RL3, none of the surveyed HHs reported any utilization post-dam. However, in the latter, the percentage of HHs reporting drinking water use in the pre-dam period was already quite low to begin with (36%). The proportion of HHs in the other four villages using riverine water for drinking purposes was also found to be negligible during the post-dam period. At the same time, there was also report of high iron content in the well water that seemed to have been aggravated after the dam obstructed river flow. I could observe dark reddish-brown colored water in the wells of some of the households located closer to the river. The water also had a strong and pungent smell indicative of high iron content and did not seem fit for either drinking or domestic purposes. There is a natural linkage between the water in the river and the ground water in areas nearer to the river, thereby fluctuating with the state of the river. The increase in iron content of the water in the riparian areas might be related to the absence of sufficient flow in the river during the post-dam period.

Table 4.5: River use and dependence of the riparian communities in the sample villages before and after dam construction.

% of HHs depending on the river for… Village Drinking water purpose Domestic purpose Driftwood collection Pre-dam Post-dam Pre-dam Post-dam Pre-dam Post-dam RL1 17 0 70 14 65 15 RL2 63 3 83 25 55 5 RL3 36 0 82 49 46 30 RR4 42 12 51 32 71 49 RR5 88 13 100 95 100 88 RR6 20 4 54 38 36 24

131

Chapter 4

On the other hand, use of the riverine water for domestic purposes did exist across all the sample villages prior to dam construction. However, the use or dependence was only occasional and not on a daily basis. A cultural practice reported was the visit to the river by women in groups before the onset of the Bihu festival for washing clothes etc. Post-dam development, use was observed to have decreased considerably across all villages except in RR5. The highest domestic use was reported by RR5 (100% HHs), which remained relatively unchanged even after dam construction (95% HHs). The decline in drinking water use mainly occurred due to the shift to other water sources (tube-wells/open wells/ water pumps), though some did report decrease in river water quality as a reason. RR5 also exhibited minor post-dam change pertaining to collection of driftwood.

Upon analyzing the primary reasons cited for the decreased dependence on river water for drinking or domestic purposes (in other villages excluding RR5), it was found that 22% in RL1, 52% in RL2, 5% in RL3, 36% in RR4, and 44% in RR6 asserted a decline in water quality following dam obstruction. These percentages include residents who have either completely stopped using the river water or reported a decreased use in the post-dam period (either for drinking or domestic or both).They complained that unlike pre-dam times, the current river water is muddy, foul smelling, murky, and blackish in color, deeming it unfit for drinking as well as domestic use. At the same time, the percentage of residents reporting a decreased dependence due to other external factors, such as shift to tube-wells/open-wells and motorized water pumps (as more convenient sources of water), were evaluated to be 11%, 9%, 46%, 29%, and 6%, respectively in RL1, RL2, RL3, RR4, and RR6. Even residents who were non-dependent during both pre-dam and post-dam periods cited similar observations of decreased water quality of the Ranganadi, and the drying up of the river during the winter season, as possible reasons for its decreased use. In case of RL3, the use of river water for domestic purposes (pre-dam dependence) was primarily limited to the occasion of Bihu. Hence, this activity, which was already rare to begin with, ceased to exist in the post-dam period. Therefore, although there was a post-dam change in dependence (as reflected in Table 4.5), it was not striking and negligibly dam-induced. The remaining HHs (across all the mentioned five villages) exhibiting a post-dam change did not give any particular reason for the decreased dependence, but it might be possible that the shift is due to the influence of extraneous factors, unrelated to the dam.

132

Chapter 4

Both the Ranganadi and Dikrong rivers originate in the sub-Himalayan region of Arunachal Pradesh and traverse down through heavy forested areas, thus bringing along large woody debris into the floodplains largely during the monsoon floods. These large woody debris or driftwood is typically caught by the downstream floodplain communities and constitute an important source of household fuel and timber (Yadama, 2013; Das et al., 2009). It is a highly risky yet traditional practice usually specific to the Mishing community forming an integral part of their diverse river-based livelihoods. In RR5, 100% of the HHs reported this traditional practice before dam construction and it declined to 88% following construction of the dam, thus indicating only a small-scale change in post-dam riverine use. One of the reasons for such high riverine use and dependence in RR5 could be their close proximity to the river where majority of the households are settled right by the river bank. Another reason could be the high proportion of the Mishing community in RR5, who are traditionally known to be river- dependent. This community also exhibits high adaptive capability to riverine hazards wherein the structure of their houses is quite distinct and suited to the ‘living with floods’ concept. Their houses are typically made of bamboo (complete with bamboo walls, tin or rice straw roofs and wooden and bamboo floors) and erected or bamboo or concrete stilts thus elevated from the ground. However, in one of the left bank villages of Ranganadi, i.e. Borbil (RL1), it was found that the flash flood of 2008 had deposited high amounts of sand in the village that resulted in elevation of the ground beneath the house, thereby decreasing the ground clearance. Since then subsequent floods after 2008, narrowly miss these bamboo houses and is now a matter of great concern to the residents.

Similar indigenous floodplain communities of Assam with historically high river dependence are the ‘Koibartta’ community, with fishing as their principal river-based livelihood, wherein, ‘Koi’ denotes ‘water’ and ‘barta’ means ‘to exist’ (Das et al., 2009).

Apart from RR5, other villages that had largely reported traditional driftwood collection were RR4 (71% HHs), RL1 (65%) and RL2 (55%).In all these three villages, there was considerable decline in the practice post-dam development, wherein 41% in RR4, 6% in RL1, and 23% in RL2 perceived that there has been a decrease in the amount of driftwood coming down the river due to obstruction in downstream river flows by the dam. They also observed that the practice has become too risky due to the high velocity of the dam-released waters. Similar reason was cited by 44% of the HHs in RR6, who

133

Chapter 4 either continued to collect driftwood post-dam or had abandoned the practice. This was despite the practice being less prevalent in the village even during pre-dam times. Increased deforestation in the upstream hills of Arunachal Pradesh had also been cited as a contribution to the decreased driftwood. Lastly, in case of RL3, driftwood collection was quite low even during the pre-dam period similar to RR6, with no notable change following dam interventions.

4.6.4 Perceptions of the downstream populations of the hydel project

Across all the sample villages, the downstream population held a general negative perception towards the upstream hydropower project and majority regarded the project to have caused more damage than benefits. Out of the total surveyed HHs, ~ 90% in RL1, RL2, and RL3 (all left bank villages) had an overall negative view for the dam. Whereas, in case of the right bank villages, perception of the dam was to some extent less negative than the left bank villages. The proportion of HHs holding a negative view in RR4, RR5, and RR6 are 49%, 67%, and 56% respectively. The difference in perception between the sample villages of the two banks could be attributed to the long-term adverse effects of flash floods and sand casting in the left bank villages consecutively in 2008 and 2009. As quoted by one of the villagers in RL2 (March, 2014),

“2008 was the year when opinions towards the upstream dam changed and it has got no support now. During the construction phase, people were unaware of what the effects would be. But now, after 10 years post-dam, people are aware that there are no benefits. This has also resulted in strong opposition against the Subansiri dam.”

Moreover the permanent geomorphological changes in the Joyhing sub-channel of Ranganadi on the left bank floodplain, which was reported to be causing more damage than the main Ranganadi in recent years (specifically post-2008), is another reason of contention between the affected villages and the upstream dam. More than 82% of the surveyed households in RL1, RL2, and RL3 claimed to have suffered in some way or the other from the upstream dam.

In the right bank villages of RR4, RR5, and RR6, at least 20%, 10%, 16% of the total HHs respectively did hold a positive view for the project and claimed to have been a benefit for the region through power production. People in these villages welcomed the overall reduction of river flows, despite dam-induced modification of the natural flood

134

Chapter 4

cycles. The view might change slightly for un-surveyed right bank villages further downstream along the Ranganadi that may have been affected by flood hazards and embankment breaches similar to RL1 and RL2.

Interestingly, irrespective of whether they have been affected by the dam or not and regardless of which side of the river they were located in, a high percentage of households across all the sample villages (> 90% in all left bank villages and RR5; > 80% in RR4) felt equally threatened by the presence of the upstream dam. RR6 was the only village wherein a comparatively lower proportion of the population (70%) felt threatened by the dam. This could be due to the pattern of spatial distribution of the households and topology in RR6, where firstly, most parts of the settled area spread perpendicularly away from the river; and secondly, the obstruction to flood waters by the embankment as well as the concrete highway leading to Arunachal Pradesh. These two factors were observed to have diminished the fear regarding probable riverine hazards, primarily the predominant fear of dam breakage and resultant downstream flood havoc as perceived by the surveyed population.

The surveyed population cited various reasons for their dislike of the upstream dam based on the riverine changes as observed by them in the past decade and personal experiences of the adverse effects. Major reasons that were cited as responsible for the downstream adversities, sense of threat from the upstream dam and hence the negative perception based on the repetitive analogous statements given by the HHs, are as follows:

1. Fear of dam breakage (with the probability of massive downstream floods)

2. Increased frequency and/or severity of floods and occurrence of flash floods

3. Uncertainty over increase in river water level and flood timing/irregular flooding

4. Fear of embankment breach

5. Sudden water release

From Fig. 4.8, it can be observed that the occurrence of flash floods and sudden water release from the upstream dam was cited the most by majority of the downstream population. The region is no stranger to floods, yet incidents such as those of 2008 where the rapid release of flood waters by the dam had resulted in downstream flash floods and

135

Chapter 4 ensuing devastation of the riparian villages, have instilled a permanent fear in the riparian communities irrespective of whether or not they were affected by the incident. Intensification of floods in downstream riparian areas by dam through sudden releases of excess water contained in the reservoir is not uncommon. Similar incidents have been

Fig. 4.8 Percentage distribution of households in each village according to the reasons cited for the downstream adversities attributed to RHEP, the sense of threat from the upstream

dam and negative perception of the hydel project. reported from multiple cases in India and elsewhere, such as the 2004 floods in the downstream areas of western Assam due to the Kuricchu hydropower project in Bhutan (Das, 2003; Thakkar, 2003), floods occurring in the River Niger due to upstream Kanji and Jebba hydroelectric dams (Olukanni and Salami, 2012). Thakkar (2003) also lists some more incidences of dam-induced floods in Indian river basins such as – 2006 floods in Tapi River due to the Ukai dam, 2008 floods in Mahanadi River due to the mismanagement of the Hirakud dam, 2010 floods in the Bhagirathi and Ganga river basins due to ‘faulty operation’ of the Tehri dam etc.

Flooding as a negative impact of hydropower projects and storage reservoirs was asserted by majority of the downstream communities (43% of the respondents) along

136

Chapter 4

River Tana in Kenya that has been fragmented by cascading dams (Okuku et al., 2015). The storage capacities of the reservoirs in the Tana River are effective only in controlling the smaller floods that result from normal annual rains. On the contrary, the reservoirs fail to contain larger floods that occur during heavy precipitation. The false sense of security provided by the reservoirs through the control over small floods caused the occupation and conversion of former floodplain areas to human settlements and agricultural lands. Larger floods spread over these areas, thereby increasing the loss of life and property (Okuku et al., 2015). Lebel et al. (2005), in the context of water governance in the Mekong Basin, believe that anthropogenic manipulations of the water resources at one place or level would invariably result in unintended side-effects elsewhere or on another level. Drawing on similar lines, Okuku et al. (2015) explain that the increased damage and loss from post-dam floods are a result of the ‘unpreparedness’ of downstream populations for such hazards, that might stem from the false sense security provided by flood-control reservoirs.

The World Commission on Dams Report (2000) extensively discusses the ambiguity of flood control by dams around the world. While some dams have been successful in controlling floods and reducing the water hazard in the downstream areas, there have also been cases where dams have increased the vulnerability of riparian communities to floods. Large dams with flood control mechanisms are designed to hold back ‘all or a portion of the flood waters’ and slowly release it over time. While the theory behind the process ensures successful flood control, in practice, it is the proper and timely operation of the dam with regular monitoring of the reservoir inflow and accurate predictions of the timing of flood peaks (based on precipitation patterns) that would ensure minimization of floods downstream and also not jeopardize the dam structure. However, most of the large dams with flood control also always have the primary function of power generation and/or irrigation; and the conflict between the various functions often lead to the exacerbation of downstream floods. Instead of slow and timely release of the inflows flooding into the reservoir, hydropower dams usually target exploitation of the situation to maximize power generation. This could prove disastrous in river basins with high precipitation patterns and especially during cloud bursts, where maximum amount of the water is released through the spillway only towards the end when the flood peaks have risen beyond containment. Another growing concern regarding the adequate management of flood control by dams is the modification of the

137

Chapter 4

hydrological conditions due to climate change, which increases the unpredictability of precipitation patterns especially cloud bursts and storm events (WCD, 2000; Totten et al., 2010; Rampini, 2016).

The World Commission on Dams report also discusses the false sense of security that flood control dams give the riparian communities according to which they encroach upon the flood-prone plains, thereby increasing the scale of damage on property and population when unanticipated floods occur. ‘Damages may therefore be larger than if floods had continued to be normal events within the range of regular experience and awareness’ (WCD, 2000). Lack of proper warning mechanisms in the downstream areas also increases the vulnerability to dam-induced floods.

Similarly, the Ranganadi hydel project lacks proper warning mechanisms for water release routines in the riparian villages especially in the remote interior areas. Warnings are aired only in the main town of Lakhimpur and the areas in the near vicinity. Hence, the residents of the actual affected areas remain unaware when water is released and many a times fail to prevent losses when the water level in the river suddenly reaches danger levels.

Uncertainty over increase in river water level and flood timing/irregular flooding was cited as another major problem in the post-dam period and a reason for the prevailing negativity towards the dam (Fig. 4.8). Although not exclusive to the villages surveyed in this study, similar observations, including the sudden water release by RHEP, late reception of warnings by villagers, and increase in damages post-construction of the dam were asserted by the riparian population along the Ranganadi, as reported by Rampini (2016). Percentage of households citing the given reason was highest in RL2 at 58% and lowest in RR4 at 17%. Floods were reported to have become more irregular and unpredictable post-dam construction, which is affecting the time of sowing and cropping (thus affecting the indigenous cropping system) and also damages the standing crop due to prolonged submergence. Although agricultural loss due to natural changes in rainfall patterns do exist, yet most of the riparian communities reported untimely floods aided by the upstream dam to be more damaging than the usual normal flood cycles. In some of the villages like RL2, residents complained that each time the sowed crop is destroyed by flood waves they have to re-sow and the process might happen multiple times and multiple waves of flood consecutively destroy the crops.

138

Chapter 4

As has been mentioned earlier, one of the main disadvantages of the floodplain communities particularly those settled in the North Brahmaputra Plains Zone is the predominant practice of rice mono-cropping especially winter or sali rice. Typical sali rice is sown at the onset of monsoon and hence is highly vulnerable to flood occurrences with prolonged inundation of the agricultural lands. Recent efforts by the government involve encouragement of summer rice (Boro rice) cultivation in the flood prone areas since the timing of cultivation does not coincide with the monsoon period. Ahmed et al. (2011) encourage the cultivation of boro rice as it is free from the risks of flood and drought and is the most productive season for rice cropping in Assam, with proper irrigation facilities. However, there are certain constraints to this with the lack of irrigation facilities being the foremost. All the surveyed villages in this study lacked irrigation facility and agriculture is primarily rain-fed. In certain cases observed in the Dikrong river basin, farmers use fuel-based pumps to extract water from the river and irrigate the fields by the river bank. The preference for sali cropping despite its vulnerability to water hazards signifies the traditional livelihood pattern in the surveyed area.

The risk of water-hazards increases during the rainy season when sudden releases by the upstream dam further elevates downstream flow in an already over flowing river. As a result, the riparian populations live in constant fear of when the water might overtop the danger level and result in flash floods with embankment breaches. Unanticipated increase in water level also puts the grazing of animals in the riverine islands at a higher risk. Incidents were reported from both RL1 and RL3, where farmers and livestock were stranded in the chapories when water levels in the river increased without warning. Another interesting observation recounted by the surveyed population was the increased velocity of the dam-released water, also one of the characteristics that helped them identify when water was being released from the upstream dam and avoid the danger. On the contrary, during winter season the river remains nearly dry, being deprived of even the minimum flow that needs to be present for supporting the riverine and riparian ecosystem.

Figure 4.8 shows that the fear of embankment breach was quite prevalent in RL1, RL3, RR4 and RR5, while it was much lower in RR6 and entirely absent in RL2. This could be owed to the fact that first four villages had experienced damages from

139

Chapter 4 embankment breaches in 2008 (left bank) and 2002 (right bank). On the other hand, RL2 is located on an island bar outside the protection of any embankment and so there is no fear of something that does not exist. RR6 is located a fair distance away from Ranganadi and is also separated by the Assam-Arunachal highway. Thus, fear of embankment breach is much less.

Another perceived threat derived from the presence of the dam upstream was - fear of dam breakage, as expressed by more than 30% households in RL2, RR4, RR5 and RR6. Geologically the Northeast region shows high seismic activity and is prone to frequent earthquakes due to the collision between the Eurasian, Indian, and Indo- Myanmar tectonic plates. The high seismicity of the region has led to questions about the safety of these dams and their degree of resilience towards high-intensity earthquakes like those of 1897 and 1950, both of which were of 8.7 magnitudes on the Richter scale. While the actual dam structure may be able to withstand earthquakes of even higher magnitude, the risks of quake induced changes in the river morphology, siltation, and aggradation of the river-bed, associated flash floods and landslides that affect the reservoir holding capacity still remain (Goswami and Das, 2003). A famous example of natural dam breakage is the Great Assam Earthquake of 1950 that triggered landslides in the mountains and formed a natural dam over the Subansiri River near the foothills. The dam blocked river flow for three to four days after which it was breached. The break resulted in a 6 m high wave down the river and devastating flash floods downstream. People were caught completely unaware (532 deaths) and many villages located by the river were washed away in the event (Sarma, 1998, 2014; Goswami, 2008). Huge socio- economic losses were incurred and the event caused many morphological changes in Subansiri as well as its tributaries, raising the bed level due to the deposition of sediment and silt from the landslide.

Likewise, the Ranganadi project falls under Seismic Zone-V (India-WRIS), but so far the project has not posed any threat. Yet, people in the surveyed villages by Ranganadi fear the repetition of the 1950 incident and therefore strongly oppose the construction of dams in an earthquake prone area with a history of landslide induced natural dam breaches. The RHEP is a man-made stable concrete gravity dam able to withstand high magnitude earthquakes; still the people living downstream lack both confidence and knowledge about the stability of such a structure.

140

Chapter 4

4.6.4.1 The Subansiri anti-dam movement and its influence upon opinion formation towards RHEP

Much of the negative opinion towards the Ranganadi hydel project seemed to be largely influenced by the agitation against the partially completed 2000 MW Lower Subansiri Project. Envisioned to be one of the largest hydel projects in India, the Lower Subansiri project has received a lot of criticism due to the large-scale environmental and socio- economic impacts that it would cause upon the downstream riparian population. At the same time, the fragile geological conditions of the project’s location in a high risk earthquake zone drew a lot of negative attention. The protests and agitation against the project finally led to halting of ongoing construction work and an expert panel with professionals from various universities in Assam and elsewhere in India was set up to conduct a comprehensive study on the viability of such a mega hydel project in an environmentally, geologically and culturally vulnerable area. While some of the suggestions of the expert panel were taken into consideration, others were ignored as is the common practice in any financially profitable infrastructure development alongside a weak environmental law setting. Nevertheless, as protests against the project still emerge from time to time, the anti-dam movement by organisations such as the Krishak Mukti Sangram Samiti (KMSS) and All Assam Students’ Union (AASU), has led to a strong negative perception against all hydropower projects in Arunachal Pradesh and thus against RHEP as well.

4.6.5 Variables influencing riverine dependence and the spatial distribution of impacts

Based on the analyses of the riverine dependence and the impacts of riverine changes upon the riparian communities, it can be observed that the possibility of an impact and the extent of it are largely governed by three variables, although there might still be some outliers. These governing variables are –

a) Proximity and location

b) Topography

c) Traditional livelihoods

141

Chapter 4 a) Proximity and location

This variable involves the proximity to the river and the location of both the homestead and agriculture lands in the riparian zone. The lateral distance of a village from the river and even within a village the distance of a particular household from the river was found to play a major role in determining the direct river dependence of the household. Proximity also affects the nature and extent of impact likely to occur both on village and on individual household level, from direct flow changes in the river and associated biophysical changes in the surrounding ecosystem. The nearer a village was to the river, more was both dependence and impact. Within the village again, the nearer a particular household’s location was to the river, higher was its dependence on the river and its ecosystem goods and services and also higher were the impacts from flow changes such as floods, timing and duration of inundation, bank erosion etc. Proximity determining dependence and vulnerability has also been reflected in the studies by Richter et al. (2010) and Vagholikar and Das (2010). People living closer to the river accessed the river more for direct use as well as recreational purposes. For example, people living nearer to the river tend to access it more for domestic water purposes as was evident in RR5 where 88% and 100% of the surveyed households accessed the river for drinking and domestic water use respectively. Practice of driftwood collection was also observed to be high in RR5 during both pre and post-dam construction. Majority of the population in RR5 lived right by the river bank separated only by the embankment (that also served as the access road to the village) and intensively cultivated the banks for vegetables during the rabi season (for both commercial and consumptive purposes). However, when it comes to impacts from natural or dam-induced riverine changes, all households living less than 2 km from the river were likely to be affected irrespective of whether they are dependent or non-dependent.

Variation in riverine dependence and impacts of river flow changes following dam closure were also observed based on the location of the agricultural lands. These variations again existed both between the different study villages as well as within a village between different households. Agricultural lands situated closer to the river bank and intrinsically connected to the riverine system were directly influenced by flow changes in the river and its sub-channels. While these alluvial farmlands benefitted from the deposition of fertile silt and nutrients during moderate flooding incidences, they are

142

Chapter 4 also more likely to suffer directly during the damaging flood events. The farmlands across the floodplain are located next to each other, occasionally separated by features such as seasonal streams or drains and artificial ponds etc. These lands are much more prone to floods, sand casting, and bank erosion problems and highly affected by channel shifting. Thus, any dam-induced flow changes in the river such as flash floods and embankment breaches would affect lands closer to the river more than those located some distance away. Therefore, even if a household is settled a safe distance away from the direct influence of the river, because of the proximity of their agricultural lands to the river were likely to be still impacted. This was observed in case of RR5, where households having farmlands across the river in the left bank were affected during the 2008 flash floods even though the same had no impact upon the homesteads on the right bank. b) Topography

This variable primarily includes natural and man-made surface land features such as river embankments, roads and railway lines, etc. From the household interviews conducted in the downstream villages, it was found that as long as a village or the section of the village was separated from the river by an embankment or an elevated road, the impact of floods and erosion and dam-induced changes in the river, was less. Until and unless there is an embankment breach or breach of road due to over flowing of river water, the villages receive adequate protection from natural as well as dam-induced flooding and erosion. RR6 is an example of one such case as has been explained earlier.

Embankments and highways are common man-made barriers that obstruct the flow of riverine water into the adjacent floodplain. However embankments as protective barriers are ‘only partially effective’ and often aggravate damage as they usually fail to contain floods that increase beyond a certain magnitude and the onslaught of frequent erosion and inadequate maintenance (due to lack of funding) weakens the structure, thus inducing breakage and breach (World Bank, 2007, pp. 43; Vagholikar and Das, 2010; Goswami, 2008). Goswami and Das (2003) further explain that in case of the Brahmaputra Basin, embankment breaching has been a ‘major cause of intensification of flood hazard in recent times’. Bhattacharyya and Bora (1997) state that poorly managed embankments may often cause more devastating and sudden flash flooding and prolonged inundation of flooded areas due to water stagnation rather than acting as effective flood- control measures. Terming river embankments as only ad-hoc measures, they state that

143

Chapter 4

embankments also have adverse effects by restricting the spread of silt over extensive areas of the floodplain and lead to drainage congestion in most areas. Embankments also restrict the river to its main course of flow and by restricting the silt and sediment to be deposited on its riparian lands instead deposit the load inside the river-bed that over a period of time leads to bed aggradation. Eventually because of risen river-bed overtopping of flood waters across embankments become easier and more frequent even during moderate flooding seasons. c) Traditional livelihoods

River and riparian dependence of downstream populations tend to vary greatly according to the dominant source of livelihood. Households where the major source of livelihood is agriculture, especially floodplain agriculture are likely to be more dependent upon the riparian ecosystem than non-farming households are. On the contrary, households those are not exclusively agricultural households but primarily engaged in non-farm economies tend to have less river dependence and are likely to be more resilient towards any biophysical changes in the riverine system and associated socio-economic impacts.

Similarly, the river and riparian wetlands are a rich source of fish for the riparian communities. The practice of fishing is not on a commercial scale yet the communities benefit from the provided ecosystem good and service and still forms a part of their traditional livelihood. Decline in fish productivity of these sources due to dam-regulated fluctuations in river flow has led to people depending upon the market for the resource. At the same time, incidents of flash floods and sand deposition affecting the individual fishponds and tanks negatively impact the livelihood of a particular household.

Thus, the discussed variables of proximity or nearness to the river and location of agricultural lands, topography and livelihood patterns, directly and indirectly, independently and in relation with one another, were likely to influence the degree of riverine dependence and impacts (environmental and socio-economic) including the lateral extent of impacts brought by biophysical and hydrological changes in the riparian system.

In conclusion, dams change the hydrology of rivers and the resultant biophysical changes often modify the inter-relationships between downstream floodplain communities and the riparian ecosystem. Likewise, the Ranganadi hydel project has

144

Chapter 4 disrupted both wet season and dry season flows in the downstream channel of the Ranganadi with modifications in the downstream channel geomorphology. More than the daily fluctuations in river flow, it was untimely occurrences of extreme events such as flash floods and sand casting over fertile agricultural lands that affected the downstream villages more. Floods are a persistent and recurrent problem in the region but majority of the surveyed population perceive that due to upstream dam regulations the floods now have become more irregular, untimely and erratic with sudden releases of excess waters retained in the reservoir. The 2008 flood also holds example of how extreme events can result in devastating and often long-term damages to floodplain agriculture and impact the traditional livelihood activities of the downstream communities. However, factors such as topography and other man-made interventions such as embankments were found to differentially moderate the intensity of such impacts across various spaces in the floodplain.

The affected population exhibited various adaptations such as agricultural extensification, livelihood diversification, shifting of cropping patterns etc. to the adverse impacts of biophysical changes on the economic and social functions. The surveyed households across all villages attributed either one or a combination of any of these adaptive strategies to compensate for the losses and reduce further risk from river-related hazards. However, despite the impacts on agriculture where complete loss of agricultural lands were also reported, a ‘farm-exit’ result was hardly observed. Instead of giving up farming entirely, majority exhibited a change in farming as a primary livelihood activity to other non-farm options such as small businesses or casual labor, while still engaged in farming as a secondary source of livelihood. Loss in farming was also compensated by adoption of multiple livelihoods.

This chapter also highlights the impact of the Ranganadi dam upon downstream fish populations in the river as well as adjoining wetlands and other low-lying areas. Although fishing was found to be more prevalent as a tradition rather than as commerce in the surveyed villages, yet low fish availability in these riparian areas has led to a decline in the practice and increased dependence on market resources. Drying up of the river and the associated riparian areas due to reduced river flows and increased sand deposition were the main reasons cited for the decline in riparian fish availability.

145

Chapter 4

Similarly, there has been a decrease in the riparian communities’ riverine dependence and use post-dam construction due to the hydrological changes in the river. There has been a decreased use of the river water for drinking and domestic purposes although the practice used to exist more prior to dam construction. Although much of the change can also be attributed to other natural processes of social change, yet the dam and its regulation of downstream flows is a contributing factor to the negative impacts and reduced dependence.

Increased flash floods, sand casting, embankment breaches etc., which have mostly been reported to have increased post-dam construction according to the riparian populations, have also led to a lot of negative perception of the upstream project. Even for households that did not show any impact or were less affected had a negative judgment about the dam as they could still witness the damages in their neighboring areas. There is a general consensus amongst the riparian populations that they all feel threatened by the presence of the upstream dam and are uncertain of when a tragedy such as that of 2008 may befall them. Many also feared that the dam might break some day and cause unprecedented floods in the area.

In the next chapter, comparisons are drawn between the hydrological and morphological changes of the Ranganadi and Dikrong rivers under the influence of the same hydroelectric project. The inherent natural and anthropogenic complications surrounding RHEP that has possibly aggravated downstream impacts in the post-dam period are discussed.

146

Chapter 5

Understanding the complexities around Ranganadi Hydel Project and broader hydropower development in Arunachal Pradesh

5.1 Introduction

Often termed as ‘flashy’ in nature, the north bank tributaries of the Brahmaputra River system originating in the Himalayas, are well known for being unpredictable with frequent channel shifting and bifurcations, given their large sediment loads and heavy monsoonal discharges (Sarma, 2004, 2014; Bora, 2004). Likewise, the Dikrong and Ranganadi are also quite unstable and dynamic, where the channels ‘often oscillate from one bank to the other’ alternately causing erosion and deposition of the two banks (SJVN Limited, 2012). Given the large silt loads that are carried down from the upper catchments and deposited onto shallow river-beds at their floodplain belts, the two rivers are continuously undergoing changes with respect to their fluvial geomorphology. Therefore, it is difficult to attribute a strict causality to the dam’s influence upon the changes occurring post-2002 where the complex natural controls of change intermix and overlap with the anthropogenic factors. Yet its influence in exacerbating the downstream impacts in the two rivers cannot be completely discounted.

The influence of the dam does not become apparent as long as the two rivers are viewed as isolated case studies. RHEP’s role becomes more evident when the timeline and order of changes in Ranganadi are analyzed in conjunction with the timeline and order of changes occurring in the Dikrong. Ideally, the two rivers that are situated in adjacent river basins with similar climatic and topographic features should exhibit similar patterns of change in their fluvial geomorphology. However, in this case, they exhibit contrasting patterns of change, that are more pronounced in the years following dam construction, thus validating the dam’s influence.

At 405 MW of installed capacity (IC), the RHEP is a much smaller project as compared to some of the mega projects (> 1000 MW of IC) being implemented in Arunachal Pradesh. Yet, it plays a significant role in the larger debate on dams in the Northeast region, predominantly in Assam and especially after the 2008 and 2017 flash floods. Following strong opposition in both states, some projects have been stalled while some have been dropped. With flawed Environmental and Social Impact Assessments

147

Chapter 5

(EIA and SIA respectively), there are multiple issues that have affected the pace of hydropower development in the upper Himalayan riparian in recent years.

Therefore, this chapter discusses three major components -

1. Firstly, the similarities and dissimilarities in the post-dam changes between the two rivers, Ranganadi and Dikrong

2. Secondly, the complexities and multiple overlapping problems surrounding RHEP (natural and anthropogenic) such as heavy water and sediment loads, failing embankments, lack of flood-plain zoning etc. that have aggravated the downstream consequence, and

3. Lastly, the current status of hydropower development in Arunachal Pradesh and broader issues therein, i.e., the downstream disadvantages.

The first half of the chapter covers the first two components, while the third component is discussed in the second half.

5.2 Discussions

5.2.1 Comparing the post-dam changes between Ranganadi and Dikrong

5.2.1.1 Hydrological changes

During the unaltered pre-dam (or pre-project) period, the median annual flow in the impounded river – Ranganadi at 120 m3/s (mean = 154 m3/s, n = 4383, 1991 to 2002), was observed to be higher than the adjacent Dikrong at 49 m3/s (mean = 86 m3/s, n = 5478, 1987 to 2002). But following water diversion, this has reversed. Ranganadi has undergone ~ 63% reduction in annual flows under dam regulations (120 m3/s to 44 m3/s), while Dikrong experienced 143% elevation (49 m3/s to 119 m3/s). April recorded the highest percentage of post-dam change in both rivers – flows reducing by 80% in Ranganadi and increasing by 319% in Dikrong. Along with March and May, the pre- monsoon months demonstrated significantly higher alteration of downstream flows in the two rivers post-2002 (Table 5.1). Similar high alteration was displayed by flow during the low rainfall months of November, December, January, and February. On the contrary, the monsoon months (or high flow months) of June to September exhibited comparatively

148

Chapter 5

lower alteration. This implies that the project has suppressed the base flow conditions in Ranganadi while elevating them in Dikrong.

5.2.1.2 Low flow alteration

Water from Ranganadi is diverted to Dikrong depending upon the inflow to the reservoir during the dry season. Uninterrupted operation of the Ranganadi power station and electricity generation even at the minimum level necessitates the transfer of water from the reservoir to the powerhouse at Dikrong, even during the dry period. Consequently, Dikrong has witnessed an increase in the magnitude and duration of low flow regimes thereby raising the threshold of minimum river flow. On the other hand, the same process has resulted in reduction of dry season flow downstream of the dam in Ranganadi. Had the powerhouse also been situated on the same river (in this case Ranganadi), then the river would have experienced increased magnitude and duration of dry season flows. Dam operations during the dry season would have involved release of reservoir-retained water onto the downstream reaches post-electricity generation, thus elevating the low flow regimes. However, this will not happen in Ranganadi, as it does not harbor the powerhouse. This could prove more disastrous for the impounded river since it jeopardizes the river’s ability to support habitat construction by the dependent riverine fauna. Movement of essential nutrients and sediments further downstream during the drier months also gets severely affected.

Hecht and Lacombe (2014) drew similar conclusions on dry season flow alteration in case of off-site generation hydropower facilities, with respect to dams in the Mekong basin. They state that dams have less impact on the seasonal timing of flow when the project has on-site generation of power, i.e., there is no water diversion between rivers. But those that divert water for power production elsewhere, even in case of run-of-river projects, substantially reduce downstream flows, ‘especially during the dry season’; while alternately increases flow in the recipient basin (Hecht and Lacombe, 2014). Similarly, the 210 MW Nam Theun-Hinboun dam upon the Nam Theun River in Lao PDR, drastically reduced the dry season flow in the impounded and donor river - Nam Theun- Nam Kading from 40-50 m3/s (under natural conditions) to 5 m3/s after diversion to the Nam-Hai (flow-recipient river) (Hirsch, 2001). Moreover, this happened despite the project being run-of-river with a small reservoir and negligible water storage.

149

Chapter 5

The observed dry flow depletion in Ranganadi suggests the lack of environmental flow23 maintenance in the river by RHEP. Based on the personal observations and those made by the local communities along Ranganadi, where the river displays a drought like situation during the dry season, it can be assumed that there are no environmental flow releases from the upstream dam, not even minimal. Moreover, given that the project is a large dam with inter-basin water diversion related off-site power generation, minimum environmental flow releases can be least expected. Assessment of environmental flows is a fairly new and developing concept in India (Smakhtin and Anputhas, 2006). Most of the methods for assessing E-flows are hydrology based and termed as minimum flows – expressed as percentage (or fraction) of the average annual flows or dry season flows or of 10-daily average flows (Tare et al., 2017). E-flows are recent additions to the impact assessment guidelines. The minimum environmental flow recommended by the Expert Appraisal Committee for River Valley and Hydroelectric Projects (under MoEF) is 20% of the average discharge over the four leanest months during the lean period and 30% of the average monsoon flow during the monsoon period on the basis of a 90% dependable year (IRG Systems South Asia Pvt. Ltd., 2015). These percentages fixed by MoEF were opposed by authorities such as NHPC, who cited that such rules were absent at the time PFRs for some of the early hydro projects were prepared. This meant projects such as RHEP that were planned in the 1980s did not require specifying the environmental flow release or adhering to them. However, the key mitigation measure for downstream impacts in a diversion associated hydro project such as RHEP, is the release of minimum ecological flow in the flow depleted river to restore the natural river habitats and existing lower riparian land-uses (Egré and Milewski, 2002).

5.2.1.3 Monsoonal flow alteration

During the monsoon season (June to September), the Ranganadi and Dikrong attain bankfull stage owing to heavy precipitation. In Ranganadi, the degree of post-dam attenuation in bankfull flows is still less as compared to the dry season flows. Similarly, in Dikrong, the degree of elevation in bankfull flows is not as high as the increase in dry

23 As per the Brisbane Declaration, environmental flows are defined as the ‘quantity, timing, and quality of water flows required to sustain freshwater and estuarine ecosystems and the human livelihoods and well- being that depend on these ecosystems’. This declaration was formed at the 10th International River Symposium and the International Environmental Flows Conference, held in Brisbane, Australia, on 3-6 September 2007. See http://www.indiawaterportal.org/sites/indiawaterportal.org/files/Brisbane_Declaration.pdf. 150

Chapter 5

Table 5.1: Differences in flow alteration between Ranganadi and Dikrong after commissioning of the RHEP

Ranganadi (R) Dikrong (D) % % Medians change change Pre-dam Post-dam Pre-dam Post-dam in R in D medians medians medians medians Annual flow (m3/s) 120 44 49 119 - 63 + 143 January (m3/s) 63 24 24 66 - 63 + 173 February (m3/s) 60 13 19 74 - 78 + 291 March (m3/s) 78 21 21 76 - 74 + 257 April (m3/s) 123 25 29 121 - 80 + 319 May (m3/s) 148 37 59 172 - 75 + 192 June (m3/s) 198 94 127 198 - 52 + 56 July (m3/s) 328 157 163 214 - 52 + 32 August (m3/s) 230 124 130 189 - 46 + 46 September (m3/s) 214 94 128 171 - 56 + 33 October (m3/s) 144 52 83 124 - 64 + 50 November (m3/s) 89 32 45 79 - 64 + 78 December (m3/s) 70 27 31 69 - 61 + 124 1-day min (m3/s) 51 6 14 28 - 88 + 100 3-day min (m3/s) 54 6 15 33 - 88 + 129 7-day min (m3/s) 55 7 15 35 - 87 + 133 30-day min (m3/s) 61 12 18 38 - 81 + 115 90-day min (m3/s) 72 21 22 67 - 71 + 207 1-day max (m3/s) 528 368 696 625 - 30 - 10 3-day max (m3/s) 487 294 538 512 - 40 -5 7-day max (m3/s) 435 249 382 439 - 43 + 15 30-day max (m3/s) 339 177 236 305 - 48 + 30 90-day max (m3/s) 290 155 161 241 - 47 + 49 Base flow index 0.32 0.12 0.16 0.24 - 63 + 50 Date of min (m3/s) 29 62 64 364 + 18 + 36 Date of max (m3/s) 185 224 187 205 + 21 + 10 Low pulse count 7 8 8 0 + 14 - 100 Low pulse duration 4 4 5 2 0 - 60 (days) High pulse count 10 6 12 12 - 40 - 4 High pulse duration 4 2 3 4 - 50 + 40 (days) Rise rate (m3/s/day) 16 4 5 9 - 74 + 70 Fall rate (m3/s/day) - 5 - 2 - 4 - 9 + 67 - 141 No. of reversals 129 109 180 188 - 16 + 4

151

Chapter 5

season flows (Table 5.1). This is so even after the addition of the maximum amount of 160 m3/s of diverted water from Ranganadi. Kellerhals et al. (1979) state that in situations (especially for alluvial rivers) where the magnitude of the diverted flows is not significantly larger than the receiving stream’s higher flood flows, then the alluvial nature of such rivers does not alter much, although there would still be modifications of channel dimensions and morphology.

The highest flood peak between 1987 and 2014 in Dikrong was 957 m3/s in 2008, which is much larger than the magnitude of the diverted flow. Even the lowest flood peak during the said period was 696 m3/s (1990), which is again much larger than 160 m3/s of diverted flow. This could explain the absence of a clearly distinct post-dam planform change in Dikrong. Although the river has widened and is more braided in the post- project period, yet planform transformations were quite pronounced in the pre-project period as well. However, the contrasting change in width and braiding between the two rivers indicates the dam’s interference. In case of Dikrong, water addition could have accelerated the ongoing natural process of geomorphic changes in the channel.

5.2.1.4 Morphological changes

The two rivers exhibit opposite patterns of temporal change in channel width, wherein Ranganadi has become narrower, and Dikrong wider. In case of both rivers, the beginning of change can be traced back to pre-dam years. In the former, the pre-dam start of channel narrowing can be explained in terms of flow obstruction by the diversion dam to fill the reservoir even prior to actual commissioning of the project. Flow may also be partially blocked and regulated to enable construction of the dam structure and the impact would increase as the project nears completion. However, in Dikrong, the change in channel width prior to flow addition is most likely due to natural factors; yet the continuing widening process in the post-dam period cannot be isolated from the influence of RHEP.

Similar to channel width, sinuosity and braiding patterns of the rivers also exhibit divergent temporal trends. Ranganadi displayed a decrease in braiding and a predominantly single-channel planform over time, while Dikrong displayed an increase in planform complexity and multi-channel braided planform. These differences in the pattern and direction of change between the two rivers, primarily post-2002, strongly suggest the dam’s involvement and the associated flow alterations.

152

Chapter 5

There was little difference between the overall pre-dam and post-dam average bankline migration rates in Ranganadi and Dikrong. However, the maximum migration rates displayed post-dam increases and specific sections along both rivers exhibited more intense bankline migration. Ranganadi recorded a decrease in the total area of erosion post-dam (2002-2014), and alternately increase in the total area of deposition. In case of Dikrong, area of erosion and deposition changed notably post-dam only for the right bank (increased erosion and decreased deposition). Besides the RHEP, other anthropogenic interventions such as embankments and dykes also played a crucial role in the reconstruction of banklines across time.

5.2.1.5 Sequence and pathway of downstream changes and socio-economic impacts

The hydrological alterations in both the rivers, exhibited by changes in the magnitude of annual and monthly flows, magnitude of annual extremes, timing of annual maximum and minimum flow conditions, etc., constitute the first order of downstream changes caused by dam construction and water diversion by RHEP. The morphological modifications in the Ranganadi – channel narrowing, decrease in braiding and channel simplification, formation of new channels and growth in the size of older partially active smaller side channels etc. constitute the second order of changes. In Dikrong, second-order morphological changes are visible during the post-dam period, yet a strong cause-effect relationship could not be established with the first-order changes. In Ranganadi, the influence of hydrological changes upon channel morphology is more evident. In both the rivers, however, bank migration is the one channel process that did not exhibit a significant post-dam change.

Flooding, flash flooding, erosion, sand-casting and related degradation of agricultural and homestead lands are the common factors impacting the socio-economic conditions of the downstream riparian populations in the impounded river. Similar observations were also reported by Rampini (2016). Degradation of agricultural lands has prompted some to shift to other livelihood options. However, instead of a complete farm- exit, adoption of multiple livelihood strategies and agricultural extensification to increase resilience to the riverine hazards and decrease total losses are the common adaptation strategies observed across the surveyed villages. Decline in riverine fisheries and other forms of river dependence ultimately indicated the socio-economic effects of environmental changes (hydrological and morphological) in the Ranganadi River,

153

Chapter 5 particularly aggravated by RHEP. It was also observed that the socio-economic effects resulted directly from the first-order hydrological anomalies in the river rather than following the sequence of first-order to third-order via second-order changes and then finally socio-economic consequences. The change in post-dam fishing patterns as explained by the surveyed population can be considered a result of third-order changes in declining fish availability caused by first-order flow changes by RHEP.

5.2.2 Floods and RHEP’s involvement

Chapter 4 discussed how the people living in the downstream areas believed that the flooding problem in Ranganadi has increased due to sudden spates of water releases from the reservoir. The unexpected rise of water level in the river along with an increase in flow velocity adversely affected the riparian populations. The problem is particularly serious because people would often navigate across the river to the opposite bank for purposes of farming, grazing their animals, or collecting fodder. The unpredictable increase in water levels due to reservoir release catches them unaware and often results in difficult situations. One such incident that was reported in a local newspaper happened on May 31, 2017, when the sudden and unwarned water release from the upstream project had resulted in the tragic death of two villagers. A resident of one of the villages in the left bank island formed between Joyhing channel and Ranganadi was ferrying his family across the river for attending a religious function (Ahmed, 2017). The hand-oared boat, which is commonly used in the rural riparian areas, capsized before they could reach the shore as the water level in the river suddenly rose along with an increase in the speed of the current due to the release of water from the upstream reservoir24.

The last flood devastation in Assam in 2017 resulted in huge losses of human lives, properties, and land, also causing internal displacement of many. Lakhimpur district particularly was one of the most severely affected. According to those residing by the Ranganadi, a primary contributing factor to the floods was the untimely release of excess water by the RHEP reservoir. The role of RHEP in aggravating the flood situation in the downstream riparian areas and causing flash flood situations was strongly highlighted by the local and state media (both print and electronic) such as Saikia (2017a, 2017b) and Matthew (2017). Failing embankments further worsened the situation. An embankment stretch of 50 m was washed away near the left bank village of Bogolijan as excess waters

24 See The Assam Tribune, 4 June 2017, p.8. The village to which the victims belonged is ‘Kulabil’.

154

Chapter 5 were released (SANDRP, 2017). Consequently, various organizations such as AASU and KMSS demanded decommissioning of the project. Following these controversies, the State Government in Assam issued directives to conduct in-depth studies of the downstream flood situation associated with RHEP operations. It called for collaboration between experts in the Indian Institute of Technology, Guwahati, representatives from NEEPCO, and other organizations such as the Asom Jatiyatabadi Yuva Chatra Parishad (AJYCP) (“IIT-G Professor asked to study Ranganadi floods”, 2017).

The RHEP has a small reservoir with gross storage of 8.10 MCM (at FRL of 567 m) and can only divert a maximum of 160 m3/s from Ranganadi to Dikrong. The ‘small pondage’ of the RHEP reservoir was also pointed out by the Ministry of Power in relation to the 2008 floods. Therefore, the project can hardly provide effective flood control as it cannot suppress the larger flood peaks arising from heavy inflows during the monsoon period. The 160 m3/s of water is also quite a small volume of water compared to the monsoonal flows in the river. Some amount of cushioning of the peak flows is possible but the reservoir has to release water (in adequate amounts) to the downstream channel to reduce the pressure on it.

Multi-purpose projects with flood moderation plans are a fairly supported idea in the Brahmaputra basin, especially with respect to the larger tributaries of the Brahmaputra. Flood control by storage dams is one of the justifications given by dam proponents and the government for hydropower development in the basin. The Brahmaputra Board explored for multi-purpose storage schemes with flood control as a primary component in the Subansiri and Siang River Basins, as early as in the 1980s (Rao, 2006). Some of the storage projects currently planned in the Brahmaputra and Barak River basins are the 2400 MW Lower Siang, 1000 MW Middle Siang (Siyom), 3000 MW Hutong, 2600 MW Kalai, 600 MW Tenga, 480 MW Kameng and 330 MW Kurung, 1500 MW Tipaimukh and 450 MW Kynshi-I dams. The Upper and Middle Subansiri HEPs were also designed to have flood cushions of 10 m and 15 m respectively besides pondage provisions. The Lower Subansiri too, has been promoted to provide flood control to downstream areas in Assam.

But as Sarma (1998) asserts, majority of the multi-purpose hydel projects in India with ‘planned and calculated storage provisions’ have often resulted in exacerbation of downstream flood situations due to release of excess water ‘synchronously with

155

Chapter 5

prolonged and intense rainfall in the basin’. The timely release of flow to reduce stress upon reservoir capacity and creating space to accommodate heavy inflows, while avoiding an overlap of the two is of utmost importance for successful flood moderation. The last-minute approach of hydel projects such as RHEP causes more trouble than providing relief. The current pattern of flow regulation by RHEP gives little flood protection to the downstream areas. It might attenuate the flood peaks, but channel modifications from extreme flow regulation, together with poorly maintained embankments accentuate the downstream flood situation despite lower peaks. Deforestation in the upper catchment also results in large amount of sediments getting transported downstream and deposited in the floodplain belt of the river, thus making the river shallower and raising the bed.

NEEPCO has repeatedly denied its role in aggravating the downstream flood situation. Both in 2008 and 2017, higher officials of the organization25 have asserted that incessant rainfall in the upper catchment caused heavy inflow into the reservoir, which far exceeded its holding capacity. Even after utilizing the maximum amount for power generation, the excess had to be released to protect the dam structure. They insisted that contrary to public allegations, the dam has provided flood cushioning to a certain extent and the downstream flood situation would be much worse if it not had been for the Ranganadi dam (Lyngdoh, 2017; “NEEPCO Not To Be Blamed For Flood In Lakhimpur: CMD”, 2017; Press Trust of India, 2017). The Ministry of Power on July 3, 2008, released a press note verifying that the flash flood on June 14 was due to heavy rainfall in the upper catchment that resulted in an inflow of ~ 428 m3/s at 2.45 a.m. and ~ 2120 m3/s at 5.10 a.m. Hence, water had to be released to the downstream for dam safety, since the reservoir has a ‘small pondage’. NEEPCO’s viewpoint regarding the dam-induced downstream impacts can probably be understood from a circular issued earlier by the company on June 2, 2006 in response to public complaints,

"...the gates of Ranganadi diversion dam may require opening from time to time ... all villages, individuals, temporary settlers etc. residing on the banks of river and other nearby areas...on the downstream of the dam to refrain from going to the river and also

25 Following the 2017 flash floods (four waves since June 9), multiple reports in the news media carried statements issued by higher officials of NEEPCO, primarily the chairman-cum-managing director, A.G. West Kharkongor and director (technical), V.K. Singh. See ‘Neepco defends Arunachal project’, The Telegraph, 19 July 2017.

156

Chapter 5

to restrict their pet animals too from moving around the river...the corporation will not take any responsibility for any loss of life of human, pet animals etc., and damage of property and others..." (Dharmadhikary, 2008).

The crux of the problem possibly lies in the mismanagement of downstream flows during flood season. The riparian areas along the Ranganadi and Dikrong are part of the eastern Brahmaputra basin that is frequently affected by water-induced hazards of floods, flash floods, bank erosion, and sand casting (Das et al., 2009). The Northeast monsoon is a well-established and well-researched system and any water development planning in the region should entail a multi-scenario flood management plan. The current manner of maximizing power generation and withholding inflow until the reservoir can hold no more, and then release of excess flows is highly likely to result in flash floods in the downstream areas. The RHEP may not have been originally built for flood moderation (which could be one of its biggest flaw), but it is time that measures are put in place to avoid the occurrences of flash floods or aggravating an already existing flood situation. The surety of heavy inflow into the reservoir during the monsoon months based on historical data should be taken into account. Even without early flood warning systems, hydel projects should have a proper flow release scheme beginning at the onset of monsoon or one that starts in the pre-monsoon stage itself, particularly designed to cope with the south-west monsoon in the Northeast.

The Assam State Disaster Management Authority (ASDMA) is responsible for disseminating flood warnings with the help of the FLEWS model developed by the Northeastern Space Applications Centre (NESAC). The FLEWS model has been installed in 14 districts across Assam to enable sending out early warning signals to the areas under the risk of flooding. However, several anomalies exist in the timely dissemination of warnings by the respective disaster management authorities and the geographical reach of such warnings. Prior to the flash flood on July 9, 2017 in Ranganadi, ASDMA did not issue any warning for the said river. Alerts were however, issued for Lakhimpur District in general over rising water levels in the Dikrong (and its tributaries) and various other sub-tributaries of the Subansiri. Yet there was no mention of Ranganadi in any of those alerts. This suggests that the flash flood incident in the river was indeed a result of sudden water release from RHEP by NEEPCO. Warnings for Dikrong were issued on 8 days

157

Chapter 5

during June and July 201726, suggesting that the river was flowing near danger level as it transported its own natural share of monsoonal flow and the added diverted amount from Ranganadi. At the same time, the diversion and obstruction of downstream flow in Ranganadi would mean that the river was not flowing near the danger level of causing a flood situation during the same dates, until water was released from the upstream reservoir. The flood reports that were published by ASDMA on July 9, 10, and 11, 201727, also did not have any mention of Ranganadi flowing above danger level. However, embankment breach of 200 m (near Amtola, Nowboicha revenue circle, Lakhimpur district) was reported on 10th July, on the right bank of the Ranganadi and again on 11th July (at Bogolijan, right bank embankment) (ASDMA, 2017).

Fig. 5.1 Hydrographs of daily flow in Ranganadi during the high rainfall period of May to September for the years 1991, 1995 and 2008. 1991 and 1995 are pre-dam years, while 2008 is post -dam.

Figure 5.1 shows the hydrographs of 1991, 1995 and 2008 during the high flow period (and flood prone period) of May to September. These three years exhibited the largest flood peaks between 1991 and 2014. It can be observed that the peak in 2008 was indeed quite sudden as the days preceding and following the peak displayed low flows in

26 See flood alerts posted by the Assam State Disaster Management Authority for details. http://sdmassam.nic.in/alerts_details_2017.html. 27 The reports give an account of the flood situation across various affected districts in Assam on a particular day as on between 3 pm to 5 pm. 158

Chapter 5

the range of 16.8 m3/s to 212.4 m3/s. On the contrary, the peaks in 1991 and 1995 displayed a gradual rise, preceded and followed by days of gradually increasing and decreasing flows. The range of May to September flows in 1991 and 1995 was also higher than the range in 2008, indicating the unsustainable post-dam pattern of obstruction in downstream flows by the project and the last minute approach adopted for release of reservoir excess.

5.2.3 Complexities surrounding RHEP

It appears that the impounded river - Ranganadi, and the people residing by it are caught in a vicious cycle of drought and flood with a complex over-lapping of multiple natural and anthropogenic stressors. The evident natural stressors are - heavy precipitation (in both the upper and lower catchments) and high sediment loads; while the anthropogenic stressors include dams, embankments or dykes, deforestation in the upper catchment and flood-plain encroachment (for human settlement and agriculture) in the lower catchment.

5.2.3.1 The problem with embankments

Embankments as flood and erosion control measures have always been deemed as ‘temporary or ad-hoc measures’, often causing more damage than relief. Much of the devastation caused by the 1988 floods in the Brahmaputra valley could be attributed to embankment breaches at 185 locations (Bhattachaiyya and Bora, 1997). Over time breaches have become a common occurrence in the basin and a major cause of intensification of flood hazards (Goswami, 2008). Even when they were first constructed in the early fifties, these embankments were placed as ‘immediate and short-term’ measures under the ‘food for work’ program (Phukan et al., 2012), while other long-term measures waited to be explored. According to Phukan et al. (2012) around 70% of the current embankment system (primarily earthen structures) along the main Brahmaputra and its tributaries were built between the mid-1950s and 1970s. But years of neglect and lack of maintenance and repair resulted in the weakening of these structures and they get easily breached. Trapping of the silt and sediment within a narrow stretch of the river over the years also caused the river-beds to rise and become shallow resulting in overtopping of flood waters. Repair works commenced only after breach incidents became a frequent occurrence and the failing embankments no longer provide the kind of protection that they initially did. For example, in the right bank villages of No. 65/68

159

Chapter 5

Grant along river Dikrong, witnessed rapid erosion from the 1980s to 2000/2002. The embankment, which also served as a road, got completely eroded away exposing the adjacent riparian areas to the river. Re-construction works began much later after land and infrastructure had already been lost to the river. Soon after the new embankment was built, it was breached again and the adjacent areas inundated by flood water. While most families out-migrated, the ones who stayed behind continue to reside right by the new embankment, still exposed to further erosion and flood calamities. Embankments also cause significant ecological changes in the riparian system by affecting the feeding and spawning grounds of the fishes (Baruah and Biswas, 2003; Dandekar, 2012). They obstruct the exchange of water, nutrients, and sediment between the river, the adjacent wetlands, and rest of the floodplain, thereby affecting the proliferation of natural fisheries.

5.2.3.2 Lack of floodplain zoning

A rise in population and lack of floodplain zoning or floodplain management plans and policies have resulted in people encroaching into these vulnerable areas and living within close proximity to the river. There has also been an increased occupation of the wetlands and other grazing spaces along the riparian areas for either agricultural purposes or human settlement. The false sense of security provided by embankments make people move closer to the banks and increase their vulnerability to flood and erosion hazards. As stated by Mitra (2004), ‘complete immunity from flood or absolute 'Flood Control' is utopian in concept and economically not viable’. It requires a combination of both structural and non-structural measures to reduce the damage. Some of the non-structural relief measures could include an assessment of available government land, which can be utilized for the resettlement and rehabilitation of families affected by erosion and displaced due to complete loss of land.

5.2.3.3 Cascade development of dams

There can be further aggravation of the downstream flood and erosion situation in the Ranganadi and Dikrong basins from cumulative impacts of multiple other projects envisioned in both the rivers in their upper catchment. The various projects coming up in the two rivers besides the existing RHEP are listed in Table 5.2. While all the projects are categorized as RoR schemes, some have storage dams that would obstruct and regulate

160

Chapter 5

Table 5.2: Hydroelectric projects planned in the Ranganadi and Dikrong basins besides the existing Ranganadi Hydel Project (RHEP) in Arunachal Pradesh

Name of IC FRL ) Type Status Project Authority HEP (MW) (m) 2 Catchment Catchment area (km

Panyor/Ranganadi River 1 Adum 25 1072 366.5 RoR Pre- SALCON-BSS Joint Panyor construction Venture 2 Panyor 21 948 494 RoR S&I JMD Power Solutions Lepa Pvt. Ltd. Middle 3 Panyor 80 494 RoR S&I Rajratna Energy * Holdings Pvt. Ltd. 4 Poma 12 RoR 5 Pareng 14.5 1421.3 119 RoR DPR Virtual Pareng Hydro Pvt. 6 Pareng-II 24 1251 226 RoR PFR -do- 7 Pareng-III 21 1115 228 RoR PFR -do- 8 Pareng-IV 24 946 315 RoR PFR -do- 9 Keyi 23 902.6 259.6 RoR - DD Hydro Power & Developers 10 Pith 13 67.22 RoR DPR Built Infrastructure

Pare/ Dikrong River 1 Par 52 848 420 RoR No environ- KVK-ECI Hydro mental Energy Pvt. clearance 2 Turu 60 612 560 RoR DPR Turu Hydro Energy Pvt. Ltd. 3 Dardu 60 400 710 RoR DPR KVK Energy& Infrastructure Ltd. 4 Pare 110 245.15 824 RoR To be NEEPCO commissioned in 2018 5 Doimukh1 80 163.2 863.38 RoR SJVN Limited 6 Papum 15 - 184.2 RoR Sonam Hydro Power Pvt. Ltd. 7 Papumpam 21 160 460 RoR Meena Entrade & Engg. 8 Senki 2 390 64.131 RoR T. K. Engg. Consortium Pvt. 9 Reysing 6 1350 87.647 RoR Geopong Enterprises Adapted from “Minutes of the 9th meeting of the Expert Appraisal Committee for River Valley and Hydroelectric Projects”, by Ministry of Environment and Forests, India, 2017; and “LIST OF CONVENTIONAL H.E. PROJECTS UNDER SURVEY & INVESTIGATION IN THE COUNTRY (As on 31.03.2018)”, by Central Electricity Authority, 2018, ANNEXURE-5.

161

Chapter 5

river flow according to the electricity generation requirements. In any case, ‘run-of-river’ does not guarantee unaltered downstream river flow28. Once completed the Ranganadi and Dikrong Basin projects would also constitute a cascade of mini and large hydel schemes that would fragment the river at multiple sites, affecting the longitudinal connectivity of the river. Large hydel projects inevitably cause a wide gamut of problems both upstream and downstream of the dam, but a series of mini projects also tend to affect the riverine and riparian environment negatively. As stated by Vagholikar and Das (2010), “irrespective of the nature of project, dams fragment rivers, breaking the organic linkages between the upstream and downstream, between the river and its floodplain.” With respect to the cascade development of dams in the Mekong River, Roberts (1995) stated that the impact of two or more projects on the same river “are likely to be multiplicative rather than merely additive,” and the cumulative impact would be much greater than a single large reservoir. River fragmentation negatively affects the movement of fish populations, and fish passages can never restore their downstream numbers to pre- dam conditions.

In case of the Ranganadi, all the future projects are located upstream of the currently operating RHEP Stage I. This suggests that the stretch of the river from the uppermost project up to RHEP Stage I, which formerly was unaffected and pristine, would get modified following impoundment. The impact might not be as significant as compared to RHEP-I, since all of the projects except Panyor HEP are below 25 MW, i.e. small RoR or mini hydel. The 80 MW Panyor HEP, located 5 km upstream of Yazali town has been planned as a storage project that would ensure increased firm power to RHEP-I located downstream. Interestingly, this project and its dam location were selected after preliminary studies revealed that the planned RHEP-II would cause significant resettlement and rehabilitation issues owing to large-scale reservoir submergence. The 130 MW RHEP-II was also envisaged to be a storage project located 5 km upstream of RHEP-I, but faced much opposition once the downstream impacts of the latter became evident. Therefore, in 200929 the Government of Arunachal Pradesh signed a memorandum of understanding (MoU) with Raajratna Energy Holdings Private Limited

28 The ambiguity and issues involved with RoR are discussed in detail later in this chapter in section ‘5.2.4.2’. 29 It is important to note the year here since it occurred after 2008 – the year flash floods devastated the downstream areas in Lakhimpur district, Assam, and RHEP-I faced significant criticism for sudden water releases from the dam.

162

Chapter 5

(REHPL), Hyderabad for preparation of Feasibility and Detailed Project reports and implementation of the Panyor HEP. As a RoR scheme with diurnal storage for peak power generation, this project was originally proposed as an alternative to RHEP-II, with a new name, at a new upstream location. Later the RoR criterion was revised to storage; returning to what RHEP-II was initially planned as, but with reduced capacity.

Figure 5.2 shows the distribution of the projects planned in Dikrong30 basin and two of them – Pare and Doimukh, are located below the RHEP powerhouse at Hoj. Both projects envisage utilizing the normal discharge in Dikrong (or Pare) and the additional flow diverted from Ranganadi. The Pare HEP has nearly been completed and was stated to be commissioned in March 2018. The Doimukh project was initially proposed to be a storage scheme with 150 MW of installed capacity. Pre-feasibility report for the same was prepared by NHPC in 2004 under the 50,000 MW hydroelectricity initiatives. However, the revised PFR prepared by SJVN Limited (entrusted by the Government of Arunachal Pradesh), changed the project into a RoR scheme with 80 MW IC (SJVN Limited, 2012).

Fig. 5.2 Locations of the hydroelectric projects (marked as red rectangles) planned and under construction in the Dikrong basin.

The RHEP-I, studied in this thesis, matches the ‘cross-watershed’ diversion type of categorization for hydropower projects made by Egré and Milewski (2002), where they

30 A similar map showing projects planned in Ranganadi basin could not be prepared as there was no information (GPS coordinates) available on their locations. 163

Chapter 5 state that such designs are adopted for increasing the energy by ‘increasing the flow in the receiving stream’ where the power house is located. Consequently, the Dikrong so far has been experiencing elevated flows downstream from the powerhouse due to flow addition by RHEP, a benefit that is to be maximized by the upcoming Pare and Doimukh schemes. However, there is likelihood that the current flow pattern would get altered again once the two mentioned projects become operational. It is hard to ascertain at this stage the direction and nature of future flow changes, but a cumulative impact assessment study in both basins is most essential and long overdue. It is also uncertain as to how the fluvial geomorphology would respond to further flow fluctuations. It is highly possible that the river would behave more erratically. So far, the Dikrong has exhibited a widening and increasingly braided pattern. This might reverse if flow obstruction and regulation by the Pare and Doimukh projects result in attenuation of downstream flows. The Dikrong might then display planform modifications similar to Ranganadi.

5.2.3.4 Lack of adequate data and baseline information

The Ganga, Brahmaputra and Barak basins fall under the classified river basins in India, with restrictions on access to crucial data (World Bank, 2007; Rahaman and Varis, 2009; Central Water Commission, 2012). The flow data comes under the ‘Official Secrets Act’ (Saikia, 2017), and users are granted access to the data only after submission of a secrecy undertaking. Even after doing so, the Central Water Commission (CWC) seldom grants the data, citing various gaps and insufficient information. An expert from Aaranyak (an NGO in Assam) stated, “The government just won’t release data to researchers, citing national security issues. That is why NEEPCO gets away always” (as quoted in Saikia, 2017b). RTI applications seeking information about dams have been refused by CWC and Central Electricity Authority (CEA) and specific dam related questions often go unanswered. Anomalies exist in the records on the prevailing number of dams and projects that have been proposed, along with mismatch between the lists provided by the central and state authorities (Bhattacharjee, 2013). There is no single platform in India where all the information regarding hydropower projects can be found and verified.

There is a notable lack of transparency with regards to NEEPCO (the project authority of RHEP) as well. The organization publishes daily data on reservoir water level (pertaining to its operational hydro projects) on its website, and the earliest data available for Ranganadi reservoir water level is December 31, 2009. Therefore, there exists no

164

Chapter 5

information on reservoir conditions and gate positions for analyzing the pattern of 2008 downstream flash floods. Information is similarly absent for the period from April 6, 2017 up to September 12, 201731. A search for those dates returns, ‘No results found!’, whereas the dates prior to April 6, and after September 12, display the data. This raises questions on the organization’s data transparency policies since the said period holds critical information of the water conditions in Ranganadi corresponding to the time of 2017 flood occurrences in Assam.

The flow data available with the state government (for Ranganadi and Dikrong) has a collection interval of once/day, signifying a major limitation in this study. Flow of a river needs to be measured over three or four times in a single day, which would then record the state of the river at various stages of flow. This is particularly essential during the monsoon season, when flow may vary with respect to the precipitation timing. For instance, water level and flow in a river may be low during the early morning hours, but may reach bankfull stage and flow near danger level towards the second half of the day, especially if there is heavy precipitation in the upper catchment. If the flow is measured only once during the morning half, then essential information gets left out for later research. This results in incomplete information regarding the hydrological characteristics of the river and studies based upon such data may lack robust analyses. Strong correlations and conclusions regarding any alterations in river flow (natural or anthropogenic induced) become difficult to establish, especially if there is a dam involved. Had there been data for multiple times a day, information that is more robust could have been provided regarding the dam’s pattern of regulation upon the downstream flow of both Ranganadi and Dikrong. Nevertheless, the available flow data in this thesis together with the morphological analyses does give adequate evidence of RHEP’s impact. Besides discharge, adequate data on sediment flow is another limitation, not just concerning Ranganadi and Dikrong, but also the Brahmaputra and its major tributaries.

Another important feature of the project is the ambiguity surrounding its categorization. According to the Hydropower Policy Report of the Ministry of Power (2008a), India, the project has been categorized as ‘storage’. At the same time, the project was mentioned as a “run-of-river scheme with a very small pondage”, in a press release

31 Searches for this data were conducted repeatedly during 2017 up to May 2018. The information is unavailable till date.

165

Chapter 5

by the Ministry of Power on July 3, 2008b, justifying the project’s incapability of controlling large floods. News media also frequently reports RHEP as run-of-river. However, given that the detailed project report of RHEP is classified and could not be accessed, verifying the facts was beyond the scope of this study.

The downstream implications of the Ranganadi hydel project is crucial as it is situated in a state that has the largest hydropower potential in the Brahmaputra Basin and the country, and the highest number of hydro projects under various stages of planning and construction. The process of hydropower development in Arunachal Pradesh plays a fundamental role in determining the future of free flowing rivers in the Brahmaputra Basin and the sustenance of the floodplain ecosystem in the downstream riparian. A review of this process also reveals the inherent problems and disadvantages of being a downstream riparian, which under the current environmental impact assessment framework in India, is recognized to be free of any direct dam impacts. The following sections discuss the hydropower development scenario in Arunachal Pradesh and the underlying problems.

5.2.4 Status of hydropower development in Arunachal Pradesh and issues therein

As per the CEA’s reassessment study, 178 large32 hydro projects amounting to a total installed capacity 62,604 MW were identified in the eight northeast states of Assam, Arunachal Pradesh, Meghalaya, Manipur, Nagaland, Mizoram, Tripura and Sikkim. As of December 2017, 19 projects with a total of 3511 MW IC are in operation, while 21 more schemes (17,434 MW IC) have been cleared by the CEA and are yet to be taken up for construction (CEA, 2017). Eighty-seven of these large projects with a potential of 50,064 MW IC are situated in Arunachal Pradesh alone (Table 5.3), and the 405 MW Ranganadi Hydel project is the only operating hydroelectric station in the state. Four schemes namely – the 2000 MW Lower Subansiri (Subansiri River, developing agency - NHPC), 110 MW Pare33 (Pare River, NEEPCO), 600 MW Kameng34 (Kameng River, NEEPCO) and 144 MW Gongri (Diran Energy Private Ltd.), are under various stages of construction.

32 Above 25 MW installed capacity 33 The Dikrong is known as Pare River in Arunachal Pradesh. 34 The Kameng is known as Jia Bharali in Assam.

166

Chapter 5

IC 0 9462 620 0 0 0 0 92 10174

No. No. projects of 0 20 2 0 0 0 0 0 2 H.E. Schemes allotted for for allotted Schemes H.E. S which on development & I held is up/yet to be up taken

IC 0 3648 299 0 0 0 0 0 3947

Schemes under under Schemes & survey investigation (S & I) of No. projects 0 26 3 0 0 0 0 0 29

IC 0 2576 295 0 0 0 0 0 2871

Schemes under under Schemes in Examination CEA of No. projects 0 3 2 0 0 0 0 0 5

IC 60 4433 0 0 0 0 0 576 5076 . Schemes allotted by States for development for by allotted Schemes States

12.pdf -

Schemes returned project to authorities of No. projects 1 11 0 0 0 0 0 2 14

wise note of Hydro Power Development, 2017”, by Central Electricity Authority, 2017. 2017. Authority, Electricity Central by 2017”, Development, Power Hydro of note wise

IC 12 0 16272 270 186 66 0 0 520 17434

Schemes cleared to and yet CEA by up for taken be construction of No. projects 1 16 1 1 1 0 0 1 21

IC 375 405 322 75 105 0 60 2169 3511

Capacity in (MW) operation of No. projects 3 1 5 1 1 0 1 7 19

a.nic.in/reports/monthly/hydro/2017/state_power

ower development in the Northeastern states of India (as of 2017). of (as India of states Northeastern the in development ower

IC 650 50064 2298 1452 1761 0 2131 4248 62604

http://www.ce

Identified Hydro Hydro Identified Power Potential as per reassessment major/ from study medium schemes 25 (above MW) of No. projects 9 87 25 11 15 0 7 24 178

- Reports, State Reassessment Potential “ Hydroelectric from

Manipur Tripura Mizoram Status of hydrop of Status 5.3 Table Assam Arunachal Pradesh Meghalaya Nagaland Sikkim Total IC = installed capacity = installed IC Adapted fromRetrieved

167

Chapter 5

A major incentive behind the intense hydropower development in the state is the generation of huge revenues from sale of hydroelectricity to other states of the country, particularly to the power-starved northern states and large metropolitan cities (Baruah, 2012; SANDRP, 2010). In the process, Arunachal Pradesh hopes to improve its economic situation. Besides, the state would receive free power from each project (fixed at a favorably negotiated percentage with the developer), thereby improving its electricity scenario. The projects would also increase connectivity within the state, as remote areas where most of the upcoming projects are located would see road construction and local area development. Hydropower projects also generate local employment and help in uplifting the socio-economic conditions around the project area through setting up of schools and other basic civic facilities. Loss of land, property, and involuntary displacement resulting from submergence and construction of project townships are the adverse socio-economic impacts. Proper resettlement and rehabilitation planning and increased economic opportunities for the local populace can counter these adverse upstream impacts. In a state like Arunachal Pradesh where most parts are still remote and people lack access to the basic amenities of life, increasing connectivity, infrastructure and local area development are of utmost importance. Hydropower development can make this possible if done in a socially and environmentally sensitive manner. Therefore, the upper riparian state stands to gain a lot from development of its water resources.

Works in the Lower Subansiri project are currently stalled owing to large-scale public protest against it in the downstream areas. This project plays a significant role in the formation of opposition against large dams in Assam, and put focus on the issue of downstream impacts in the country. In Chapter 4, I had discussed how the historical geologic events in the Subansiri basin and the protests against the Lower Subansiri project influenced public opinion towards the Ranganadi hydel project. At the same time, the current flood experiences in the downstream riparian areas of RHEP, has to some extent, increased public resentment and resistance in Assam (particularly in the flood affected Lakhimpur and Dhemaji districts) to the Lower Subansiri and other large and mega hydel projects in Arunachal Pradesh. “The release of water has become a rallying point for the people of Lakhimpur against the project. It has also fuelled fears against construction of big dams” (Singh, 2017).

168

Chapter 5

5.2.4.1 Hydropower development in the Subansiri basin

The Subansiri basin was identified to have a probable hydropower potential of 12248 MW of installed capacity (Government of Arunachal Pradesh, 2008). Eighteen projects with IC > 25 MW and 10 small HEPs on the mainstem of the Subansiri River and its tributaries have been planned to realize this potential. Almost all the projects, except the Middle and Upper Subansiri hydel projects, are stated to be RoR, although with dams and reservoirs (i.e., with pondage) to meet the peaking power requirements (IRG Systems South Asia Private Ltd., 2015). The two excepting ones are to have 15 m and 10 m flood cushions respectively, which would provide additional flood regulation in the basin. A list of these projects is included in Appendix A1.

Initially, a single 4800 MW Subansiri dam was planned in the basin, and the detailed project report (DPR) for the same was prepared by the Brahmaputra Board in 1983. Opposition by the government of Arunachal Pradesh citing large scale submergence and displacement in the upper riparian (particularly submergence of the Daporijo town) resulted in its segregation to three projects – Lower Subansiri (over River Subansiri at Gerukamukh), Middle Subansiri (over River Kamla at Tamen and Upper Subansiri (over River Subansiri at Menga) (SJVN Limited, 2012). Later additional projects were also formulated to be developed in a cascade manner. The 2000 MW Lower Subansiri is the downstream most project located at Gerukamukh (near the border between Arunachal Pradesh and Assam), ~ 40 km upstream of North Lakhimpur and ~ 90 km upstream of the confluence with Brahmaputra. Hence, this project essentially controls the magnitude, timing, and pattern of flow that would finally enter the Assam floodplain.

As per the cumulative impact assessment study report submitted to the CWC in 2015 by IRG Systems South Asia Private Ltd., the Lower Subansiri project involves a concrete gravity dam of 116 m height (from the deepest foundation), FRL of 205 m and 1365 MCM of gross storage at FRL that would submerge an area of 33.5 km2. It has a surface powerhouse with eight units of Francis turbines. A continuous flow of 240 m3/s will be maintained downstream of the Lower Subansiri dam, since it has a toe powerhouse whereby one unit of turbine will have to be continuously run. Therefore, flow series computed for the river for the non-monsoon period involves release of 240

169

Chapter 5

m3/s for 20 hrs and 2579 m3/s for four hours daily (based on 4 hour peaking release scenario35).

The impact assessment report lays much stress on the fact that the monsoonal peaking release pattern of the projects in the Subansiri Basin particularly that of Lower Subansiri, would have insignificant impact upon the flow regime of the Brahmaputra River. The natural average monsoon discharge in the latter is ~ 20,000 m3/s, while the non-monsoon average (at Guwahati, 328 km downstream of Lower Subansiri) is ~ 5300 m3/s (IRG Systems South Asia Pvt. Ltd., 2015). Accordingly, a 3- hour peaking release would end in discharge increasing to 5440 m3/s and a 4-hour peaking release would increase discharge to 5540 m3/s. The simulated flow calculations in the report assert that although there might be an increase in downstream flow due to non-monsoon peaking release, it would be less than 200 m3/s. The corresponding increase in water levels would be limited to 5 to 12 cm with reference to the natural condition. Beyond the first 40 km downstream from the dam, the impacts would be less and negligible after the confluence with Brahmaputra (IRG Systems South Asia Pvt. Ltd., 2015).

The first 40 km downstream from the Lower Subansiri dam, however, would suffer significant flow alterations. Water level because of peaking release would fluctuate between 1.5 m to 2 m daily. A jump from 20 hours of 240 m3/s to 4 hours of 2579 m3/s, particularly during the non-monsoon period, is not a minor change. The 1.5 to 2 m fluctuation in water level is also not a simple change. Such a high degree of flow regulation and alteration would have significant repercussions, with higher chances of increased erosion along the banks, most likely from slumping. Bank slumping is a common phenomenon witnessed during the falling stage in the Brahmaputra River and its tributaries. As the water rises (corresponding to a flood stage), it seeps into the banks and the sediments become highly saturated, also increasing the pore pressure (Palaniappan, 2004). When the water goes down suddenly, the cohesive forces between the water molecules and change in pore pressure result in rapid withdrawal of the water trapped in the banks. The saturated sediments often liquefy and move along with the lateral movement of the water into the channel, which ultimately creates caves in the banks, fracturing of the overlaying bank material into blocks that finally collapse by tilt

35 A 3-hour peaking release would involve release of 240 m3/s for 20 hours and 2579 m3/s for 3 hours daily (IRG Systems South Asia Private Ltd., 2015). 170

Chapter 5

(Palaniappan, 2004; Sarma and Phukan, 2006). Channel migration would eventually become quite erratic.

This is a major concern for the floodplain communities living downstream, who already face problems of floods and land loss due to erosion. The projects may have negligible impact beyond the 1st 40 km. However, one needs to remember that the first 40 km of downstream riparian are not wastelands and uninhabited that can become collateral damage. These are extremely fertile lands, which support a wide array of natural and agricultural ecosystems. There are different ethnic communities living within the 1st 40 km, people who still practice traditional forms of agriculture. Their livelihood patterns depend largely on the natural goods and services provided by the riverine and riparian ecosystem sustained by the Subansiri River. Annual floods regularly affect these people and are highly vulnerable to daily and seasonal fluctuations in river flow. Moreover, peaking power releases from RoR schemes with large storage dams subject the downstream river to alternate flood and drought conditions and daily fluctuations that can wash away the riverine fish, harm their breeding grounds and destroy downstream ecosystem (International Rivers, 2017).

None of the cumulative impact assessment and carrying capacity reports for the Subansiri, Lohit and Dibang basins discusses the socio-economic impacts of dam construction on the downstream communities of the lower catchments. Although the floodplain areas of Assam fall within these basins, the impact on the riparian communities, even the ones living within close proximity to the rivers, do not merit any mention in the assessments. Uncertainty regarding the downstream limits of such impacts does exist, which makes their study prior to dam construction challenging. However, the lateral riparian areas corresponding to the longitudinal extent of the river up till which water level fluctuation would be significant (for instance 40 km in case of the Lower Subansiri project), can be regarded as the downstream limit for socio-economic and ecological changes.

The initial oppositions against the Lower Subansiri project began after the public hearing process in September 2001, during which non-governmental organizations such as the Aaranyak and Kalpvriksh Environmental Action Group, criticized the process. According to Choudhury (2014), the Lower Subansiri project was the first large hydro project to be taken up for construction in the region and the local communities were

171

Chapter 5 unfamiliar with the legalities and procedures of the public involvement process. Allegations against the developer include flaws in the EIA report which undermined the actual impact of the project and did not contain sufficient baseline information; building of project related residences prior to obtaining statutory environmental and forest clearances. Even after the project received conditional environmental clearances following reassessment of the impacts and that no other project would be constructed upstream of the Lower Subansiri, it failed to achieve public confidence and acceptance. Fresh oppositions cropped up that gained momentum under the aegis of influential semi- political organizations such as the AASU and several other civil society and non- governmental agencies. Mass social mobilization against the project, demonstrations, road blockades (that prevented the transport of constructions materials to the project site) and petitions drew extensive attention towards the Lower Subansiri Project (Choudhury, 2014). A tripartite meeting was held between the Government of Assam, the project developer (NHPC Ltd.) and AASU on December 8, 2006, that resulted in the formation of an expert committee (comprising of academicians from Gauhati University, Dibrugarh University and Indian Institute of Technology Guwahati) to carry out a comprehensive downstream impact assessment study of the Lower Subansiri project. Choudhury (2014) termed this turn of events as a ‘watershed moment in the conflict over the LSP’. While the expert committee put forward many recommendations, such as lowering of the dam height, sustainable moderation of downstream flows to minimize the downstream dam impacts etc., most were contested and overlooked by the project developer.

5.2.4.2 ‘Run-of-river’ – a misleading categorization

A majority of the projects planned in Arunachal Pradesh and elsewhere in the Brahmaputra Basin have been categorized as ‘run-of-river’ (RoR). As per the pre- feasibility reports (PFR) prepared under the 50,000 MW Hydroelectric Initiative (sanctioned in March 2003), 34 schemes out of a total of 4236 in Arunachal Pradesh were RoR hydel schemes (see Appendix A2). The installed capacity of these 32 RoR schemes ranged from a minimum of 30 MW (Tarangwarang HEP, East Kameng district) to a maximum of 4000 MW (Etalin HEP, Dibang Valley district). With respect to the dam height, the projects ranged from 17 m to 241 m (110 MW Pakke and 2600 MW Kalai projects, respectively); full reservoir level (FRL) ranged between 165 m (150 MW

36 These 42 schemes did not include projects such as the Upper, Middle and Lower Subansiri.

172

Chapter 5

Doimukh storage37) and 1950 m (700 MW Oju I); submergence area varied between 1.7 ha (30 MW Tarangwarang) and 3764 ha (600 MW Kameng dam) (Central Electricity Authority, n.d.). The wide variation between the installed capacities, dam height, full reservoir level (FRL), live storage characteristics etc. of these projects that have all been assigned RoR categories raises questions upon the universal definition of RoR. Generally, RoR schemes (with minimal or no storage) are regarded to be more sustainable compared to large and mega hydel projects due to their low submergence area, marginal displacement, and lesser ‘ecological footprints per MW’ (Kumar and Katoch, 2014). Small RoR projects are closest to maintaining the natural flow in the river as opposed to storage-based projects (Egré and Milewski, 2002).

Run-of- river signifies the generation of hydroelectricity without or with minimum water storage, using the river’s usual flow, while maintaining the balance between inflow and outflow of water, sediments, and nutrients (Anderson et al., 2015, Roberts, 1995). The definition, design, and size of these projects vary widely across the world. RoR schemes are primarily justified for being environmentally and ecologically benign, with low hydrological changes and overall impacts as compared to large storage projects (Anderson et al., 2015; Bilotta et al., 2016). Theoretically, no dam involving storage and flow regulation should be considered run-of-river; but the reality differs (Seth, 2014).

In India, a RoR power station is defined as, “a power station utilizing the run of the river flows for generation of power with sufficient pondage for supplying water for meeting diurnal or weekly fluctuations of demand. In such stations the normal course of the river is not materially altered” (Bureau of Indian Standards, 1989, p.2). Firstly, this definition lacks clarity in allowing for water storage without specifying the upper limit of ‘sufficient storage’. Hence, there are projects that range from 0 MCM (eg. 150 MW Ringong, Upper Siang district) to 1177 MCM (600 MW Kameng dam, East Kameng district) of live storage capacity. Diversion tunnels in RoR projects over the Teesta in Sikkim would result in more of the river flowing through tunnels inside the mountains than its natural course (International Rivers, 2016). It could be considered that hydropower projects in India, irrespective of their size and scale of possible impacts, are

37 The Doimukh hydel project was initially slated to be a 150 MW storage project with 67.21 Mcum live storage, 165 m FRL and 565 ha of submergence area (as per the PFRs prepared under the 50000 MW initiative). However, according to the PFR prepared by the project authority (SJVN Limited), the installed capacity was reduced to 80 MW, FRL increased to 168 m, water spread at FRL reduced to 97.36 ha and live storage at 3.89 Mcum.

173

Chapter 5

being labeled as RoR to reduce opposition against them and mislead the common man into accepting these schemes.

Roberts (1995), in relation to hydropower dams in mainstream Mekong, considered that a more apt description for run-of-the-river dams would be ‘ruin-of-the- river’. Dam proponents increasingly advocate RoR projects as being environmentally and socially benign (Huber and Joshi, 2015). They are portrayed as better alternatives to large dams due to smaller reservoir submergence, lower displacement of communities and lower downstream flow reduction, which Seth (2014) believes has given a ‘fresh lease of life to many dam builders around the globe’. In India, particularly, majority of the hydropower projects, especially those coming up in the Himalayan region are being promoted as RoR projects.

They are emphasized as benign since the total downstream flow over any 10-day period would not change and remain similar to the natural pre-dam conditions (Vagholikar and Das, 2010). However, this justification seems ambiguous as most of them involve large dams (dam height > 15 m) and significant water storage that would unarguably result in considerable downstream flow alteration. For example, the 2000 MW Lower Subansiri dam and 1750 MW Lower Demwe38 project have dams that are 116 m and 163 m high respectively. In a letter in September 7, 2009, to the then Indian Prime Minister Manmohan Singh, a group of citizens and civil society groups from Assam, claimed the promotion of RoR projects based upon their environmental benignness as a false assurance and an ‘ecological lie by the government’. Similarly, Vagholikar and Das (2010) stated, “It is clearly misleading to universally label RoR projects as socially and environmentally benign projects”. Hence, marked difference exists between the text-book definition of RoR and the actual operation and downstream flow regulation by these projects. The cumulative impacts of these projects would be even greater since a majority of them is being developed in a cascade manner. For instance, the Lohit River has a series of seven hydel projects on its main stem and an additional five on its tributaries.

38 Initially the Demwe hydroelectric project on the Lohit River was slated to be a single storage project of 3000 MW. But it was split into two RoR schemes – the 1800 MW Upper Demwe and 1750 MW Lower Demwe projects, to prevent the submergence of the Kamlang Wildlife sanctuary on the left bank of the river and ensure sustainable environmental development (CISMHE, 2009).

174

Chapter 5

Kumar and Katoch (2016) have also highlighted the ‘unsustainable construction and operational practices’ of large RoR projects in the western Himalayan region. Their study concluded that not all the environmental impacts of small hydropower projects (with reference to the RoR ones) are ‘small’ in comparison to large hydro schemes. However, the amount of literature present on the impacts of RoR projects, above all their downstream impacts, is quite limited (Csiki and Rhoades, 2010), as most of the focus still lies upon large dams.

5.2.4.3 The involvement of private players and controversies therein

Vagholikar and Das (2010) discuss the accelerated development of hydro projects in the Northeast region, more so in Arunachal Pradesh, with the advent of private hydropower companies. The then Minister of Environment and Forests (India), Mr. Jairam Ramesh, used the term ‘MoU virus’ to refer to the rapid rate of project grants by the government of Arunachal Pradesh to private companies (“Arunachal Pradesh hit by MoU virus”, 2009; Vagholikar and Das, 2010). The New Hydropower Policy formulated by the Ministry of Power in 2008, explicitly states the need for increasing the role of the private sector in hydropower development. According to the report, the public sector alone is not adequate to develop the vast untapped hydro potential of the country. The respective state governments would allocate the potential sites for hydro development within each state. As of 2008, 30 projects had been allocated by the state government in Arunachal Pradesh to private sector developers and were yet to be taken up for construction. In Sikkim, this number stood at 22, while 3 projects were already under construction (Ministry of Power, 2008a). The state hydropower policy formulated by Arunachal Pradesh in 2008 clearly states one of its objectives as “to accelerate the pace of hydropower development through participation of both the Central Public Sector Undertakings and private power developers, as also by formulating Public Private Partnership.” (Government of Arunachal Pradesh, 2008, p.3). According to the policy, the hydropower projects would be built by the private developer on ‘Build, Own, Operate and Transfer (BOOT)’ basis, with no liability on part of the Government of Arunachal Pradesh. The policy entails that the developer shall operate the project for a period of 40 years, following which it would be transferred back to the state government ‘free of cost and in good condition’. Whether the reality matches with the written policy is yet to be ascertained. No less than 12% of the power generated would be supplied freely to the state and it would have equities of

175

Chapter 5

minimum 11% in a privately developed project39. Another 1% free power would be provided to the Local Area Development Fund and credit 0.1% of the project cost to the State Government as ‘Project Monitoring Evaluation and Co-ordination (both technical and financial) fees’ (Government of Arunachal Pradesh, 2008).

Moreover, dam developers have to pay a non-refundable ‘Upfront Premium’ including the processing fees, even before the project receives any clearance. The minimum is 1 lakh per MW for a project with capacity between 25 to 99 MW. Consequently, the state government has already procured huge amount of revenues from such upfront payments. For instance, it received 930 million INR in 2007-2008 from Athena Energy Ventures Pvt. Ltd. as an upfront premium for constructing the 1750 MW Demwe Lower hydel project40, although the techno-economic clearance and environmental clearances were received a year later in 2009 and 2010 respectively (Bhattacharjee, 2013). Because of this, Bhattacharjee (2013) claims that developers exert pressure over the state government for faster clearances of the project since delays incur losses and the state government in turn pressurize the central authorities. In the process, environmental and socio-economic impact assessments and considerations get neglected and bypassed. The project developer is, further, penalized if the project is not commissioned within the stipulated time-period except under circumstances that lie beyond the control of the developer (Government of Arunachal Pradesh, 2008). The minimum penalty to be paid to the state government is Rs. 10,000 per MW per month for projects ranging between 25 MW and 99 MW IC; and a maximum of Rs. 60,000 per MW per month for projects of 3000 MW IC and above. The penalty necessitates faster clearance of projects and might explain the rapid rate at which projects in the eastern Himalayas have received clearances despite having notable shortcomings. Though substantial risks are involved given expensive premiums and penalties, the private developers too hope to reap huge benefits from generation and sale of hydroelectricity during the peak load period in the open market, termed as ‘merchant sales’ (Vagholikar, 2011).

39 The equity participation would be governed by the provisions of the Indian Company’s Act (Government of Arunachal Pradesh, 2008). 40 An RTI filed in 2008 returned documents from the Department of Hydropower, Arunachal Pradesh, revealing the non-refundable premium paid by Athena Energy Ventures Pvt. Ltd. nearly a year before receiving any formal clearance for the Demwe Lower project on Lohit River (Bhattacharjee, 2013).

176

Chapter 5

Criticisms have stemmed furthermore from concerns regarding the experience of the private developers with respect to building hydropower projects. Prior experience in the industry does not seem to be an important criterion in their selection (“Reservoir of dams”, 2008), whereby technical and financial capabilities are being given preference (Government of Arunachal Pradesh, 2008). Finally, although the state can profit economically from the hydropower projects, there is uncertainty regarding successful and holistic socio-economic development. This is because the private developers, especially the newcomers, may lack credibility with respect to their social and environmental sensitivity towards dam impacts.

Given the vast number of power projects planned in Arunachal Pradesh and the complexities and vulnerabilities of the region, the Ministry of Environment and Forests had commissioned cumulative impact assessment and carrying capacity studies for all major river basins in the state (Ghanekar, 2017). In 2017 seven projects in the Kameng basin, namely Kameng-I, Kameng-II, Pakke, Seba, Pasar, Bichom stage-I and Bichom-II, amounting to a total capacity of 2337 MW, were recommended41 to be ‘dropped’ by MoEF due to their close proximity to Pakke Tiger Reserve. It was also suggested that no further project (even lower than 25 MW or outside the purview of EIA Notification, 2006) should be constructed in the entire Kameng basin (Bansal, 2017). Similarly, the cumulative impact assessment and carrying capacity study for the Lohit River Basin recommended the abandonment of the Hutong hydroelectric project Stage 1. The cascading dams would leave only 19.1 km of free flow river stretch out of a total of 109 km. Moreover, the dam location of Hutong project is geologically unstable and bears both the Mahaseer and Snow Trout fish species, hence dropping the concerned project would increase free flow stretch to ~ 50 km, allowing migration of the fish species (WAPCOS, 2016).

41 The recommendation by MoEF was made based on the ‘Cumulative Impact Assessment and Carrying Capacity Study of the Kameng River basin by WAPCOS (Water And Power Consultancy Service Limited) (Ghanekar, 2017; Bansal, 2017).

177

Chapter 5

5.2.5 Environmental Impact Assessment of hydropower development in India - disadvantages of the downstream riparian

Environmental Impact Assessment (EIA) was formally and legally established in India in 199442 to obtain Environmental Clearance (EC) for large infrastructure projects that are likely to cause significant environmental impacts, under the Environment Protection Act of 1986 (Panigrahi and Amirupa, 2012; Erlewein, 2013). Under the currently followed EIA Notification of 2006 (and its amendments), clearance is mandatory from the Ministry of Environment and Forests (MoEF), India for 39 types of infrastructure projects, including hydropower projects above 25 MW installed capacity. There are four stages prior to environmental clearance – screening, scoping, public consultation, and appraisal (MoEF, 2006).

However, the Environmental Impact Assessments for hydropower projects in India are flawed with serious gaps and deficiencies, whereby developers often do partial assessments that are ‘inadequate and misleading’ to acquire environmental clearances (Panigrahi and Amirapu, 2012; Hill, 2017). The overall EIA system in India has not yet evolved satisfactorily and is used more as a technical exercise and project justification than project planning tool in its present form (Baruah, 2012; Panigrahi and Amirapu, 2012). The public consultation component of the EIA involves sharing of information with the people that are likely to be affected by environmental impacts from hydropower projects (the ‘plausible’ stakeholders) and take into account their concerns and views (MoEF, 2006)43. It is a crucial stage, particularly in a democratic environment, where the authorities have an opportunity to gain confidence of the general public and promote the beneficial outcomes of the project. A similar opinion is put forth by Glucker et al. (2013, p.104) stating, “Public participation in EIA is commonly deemed to foster democratic policy-making and to render EIA more effective”. The World Commission on Dams (2000) identifies prior, free, and informed consent of the public as a keystone strategic priority in dam building. However, the actual reality is contradictory where public consultations are ‘shoddy’, insufficient and lack critical information (Bhattacharjee, 2013). Until recently, the notifications for public hearings were not circulated in a proper

42 As the ‘ENVIRONMENT IMPACT ASSESSMENT NOTIFICATION S.O.60(E), dated 27/01/1994’, by the Ministry of Environment and Forests, India. 43 For details, see the Environmental Impact Assessment Notification, 2006, published by the Ministry of Environment and Forests, India, and dated 14 September 2006. http://envfor.nic.in/sites/default/files/so1533_4.pdf. 178

Chapter 5

manner (Rampini, 2016). Even in the case of the Ranganadi hydel project, the advertisements were not circulated properly that resulted in poor attendance and the hearing was rather brief in nature (Bhattacharjee, 2013).

After multiple amendments and revisions, public hearing (the public involvement process) as per the EIA notification of 2006, still takes place after a project has cleared the screening and scoping stage and the EIA report has been prepared. Moreover, the present notification allows only the locally affected populace to participate in the hearing process. Whereas, according to the 1994 EIA guidelines earlier, anybody could participate in the hearing process and submit written concerns (Choudhury, 2014). Another serious flaw in the new notification is the provision for project developers to bypass the public hearing process if there is ‘inadequate law and order situation in the project area’ (Choudhury, 2014), whereas in the past, public hearings were mandatory.

Public hearing being restricted to the locally affected populace automatically excludes the participation of downstream population in the planning process and seriously undermines the existence of downstream dam impacts, especially the socio-economic effects of impoundment. Bhattacharjee (2013) stated that according to the officials of the Dam and Research Wing of the Central Water Commission (CWC, India), dams affect river flow only in the upstream mountainous terrain and not in the downstream, thereby not needing public hearing in the downstream areas. The author asserted that this kind of ignorance by one of the principal agencies responsible for water resources development and overseeing dam sanctioning and regulation in the country is a serious issue and loophole in the system. At the same time there is no compensation for land loss in the downstream region since the land is not acquisitioned by the builder for construction of the dam as per the comments by one of the dam builders in India (“Reservoir of dams”, 2008).

Even where cumulative impact assessments are carried out for the entire river basin, the downstream impacts lack robust assessment and face longitudinal and lateral ‘coverage’ limitations (Hill, 2017). As per the EIA guidelines, the area of project influence varies according to the studied environmental component and determined during the scoping stage. The project influence area for catchment area treatment plans and land environmental studies extends from the directly draining catchment area up to the dam site. Similarly, biological and socio-economic studies are carried out in areas that

179

Chapter 5

would be directly submerged due to reservoir creation, land diverted for project components such as construction of townships, roads, identified land for rehabilitation and resettlement, muck and debris disposal, etc. Water related studies extend from the submergence area upstream of the dam site up to the powerhouse location. Finally, other sensitive locations likely to be impacted by the hydropower project are defined by a 10 km radius from the project boundary that includes the reservoir, dam, powerhouse and tailrace tunnel sites.

Thus, the Indian EIA guidelines limit the downstream environmental impact assessment for hydropower development to 10 km only. This 10 km does not include socio-economic impact assessment of communities settled below the dam. This contradicts the ground reality where significant changes in the riverine and riparian system continue further downstream. The off-site changes witnessed in the Ranganadi, approximately 30 km away from the RHEP dam, establish that the longitudinal reach of dam impacts can extend well beyond the 10 km limit. One of the key modifications that could be implemented regarding the downstream EIA limit is that, it can be demarcated on a case-by-case basis and designed according to each individual project and the extent of its probable downstream impacts. As suggested earlier44, with respect to the Lower Subansiri project, the 1st 40 km up to which significant dam-induced flow fluctuations have been projected, could be distinguished as the downstream extent for ecological and socio-economic impact studies. Similar assumptions can be considered with other hydropower projects where the extents up till which significant flow fluctuations are likely to occur can be demarcated as the downstream limit for ecological and socio- economic impact studies for a particular project.

Additionally, the downstream state has no say in the decision-making, planning or management process of hydropower development occurring in the upper riparian. The public hearings of these projects are restricted to the areas where they are situated, i.e., to the project areas. If one looks at the larger picture between China, India, and Bangladesh, then India is caught in a similar fix with China as Assam is with Arunachal Pradesh. It is not justifiable to draw parallels between the highly complex international and domestic scenarios. Nevertheless, the basic underlying obstacle in both cases is the downstream riparian having limited negotiating power with the upstream riparian. Bulk of the hydro

44 In section, ‘5.2.4.1 Hydropower development in the Subansiri basin’. 180

Chapter 5

potential lies in the upper Himalayan state (and hence the rapid river resources development), while Assam has been assessed to have a potential of only 650 MW of total installed capacity from medium/major schemes. The latter has nine identified hydropower projects, with three schemes amounting to 375 MW currently under operation. These are the 225 MW Kopili (commissioned in 2003), 100 MW Karbi Langpi (2007) and 50 MW Khandong (1984). Although Assam stands to be greatly affected by the projects in the upper riparian, it has neither any say in the planning process, nor would receive any form of compensation to losses incurred from dam-induced riverine calamities. It is this lack of space to participate in the development process or voice concerns during the planning stage that causes downstream communities to resort to various forms of protests. Popular civil society movements and protests are often the only ways in which downstream states can highlight the negative impacts of dams. The agitation against the Lower Subansiri project is an example of this.

The rate at which the Arunachal Pradesh Government has signed MoUs with hydro developers suggests that it is least bothered about the downstream problems of floods and navigation in Assam. There exists a mismatch between benefits and losses between the two riparian, which increases the possibility of trans-boundary water conflicts. Even during the 2008 flash floods in Ranganadi, the downstream affected population alleged that NEEPCO was pressurized to release water from the dam by the locals around the project area in the upper riparian. Heavy inflows into the reservoir had increased submergence area and affected the locals there, causing them to pressurize the authorities to open the gates and release the water downstream. Some of the interviewed households in the Ranganadi villages asserted this during the field survey. Conflicts relating to land disputes at the border belt near Lakhimpur district have occurred in the past between the two states. Trans-boundary water issues increases the risk of future inter-state conflict.

Not just Arunachal Pradesh, this vulnerable floodplain is affected by flash floods caused by reservoir releases from dams in the neighboring country of Bhutan as well. Sudden flood releases from the Kurichu dam in Bhutan had resulted in flash floods in the trans-boundary Manas and Beki rivers in Assam during multiple years in 2004, 2007, 2008, 2009, and 2016 (SANDRP, 2016). Downstream populations assert significant morphological changes in the rivers owing to frequent dam-induced flash flood

181

Chapter 5

occurrences. The rivers have become much shallower and wider, with huge tracts of former fertile lands laden with silt and sand (Duarah, 2015). The most affected are the foothill areas at the Indo-Bhutan border inhabited by economically poor communities with forest and river dependent livelihoods.

The risk of downstream floods in Assam from dams in Bhutan increases more due to the formation of landslide triggered artificial dam and lakes above the dams in the latter. Additionally, the Himalayan Rivers are glacier fed and prone to high discharges from Glacial Lake Outburst Flood (GLOF) incidences. In 2004, unprecedented floods and inundation by the Manas and Beki rivers occurred in the Barpeta and Nalbari districts (Lower Assam) due to unannounced water releases from the 60 MW Kurichu power project. The primary cause of this release was the bursting of a landslide-dammed artificial lake (Tsatichu Lake) that had formed 30 km upstream of the project (“Floods affect over 30,000 in Assam”, 2016). Inflows had to be released downstream to protect the dam structure but resulted in devastating floods in the floodplains of Assam, affecting more than 20,000 population and inundating large tracts of agricultural and forest lands. A considerable part of the Manas National Park, a UNESCO World Heritage site, was also impacted negatively (Kalita, 2016).

As is the norm commonly observed, the Kurichu project authority denied downstream allegations (“Floods affect over 30,000 in Assam”, 2016), citing that water released from all dams in Bhutan are normal discharges observed naturally in the river (Duarah, 2015)45. However, as discussed earlier, the operation and design of a run-of- river project differs notably in theory and practice. Although demands have been made by various organizations and the general public to the state and central governments in India initiating dialogue and adoption of early flood release warnings from Bhutan, there has been limited progress.

Without a doubt, hydropower projects cause as much negative hydrological, environmental, and socio-economic impacts downstream of the dam as in the upstream. Therefore, in case of trans-boundary river basins there should be space for the

45 The project authorities and foreign ministry of Bhutan assert that all the hydropower projects in the country are run-of-river with no significant reservoir storage. Opening of the reservoir gates to flush out accumulated silt and logs behind the dam require prior permission from the Central Water Commission in India. For details, see ‘Flash floods are burying the lands on the India-Bhutan border in silt’, UNDERSTANDING ASIA’S WATER CRISIS, thethirdpole.net, 20 November 2015.

182

Chapter 5

downstream states in the decision-making process in the domestic environment. The downstream riparian should be recognized as a stakeholder in hydropower projects. In the international arena, riparian countries sharing trans-boundary river basins should strive for better dialogue, consensus over sharing of water resources through sustainable practices as well as sharing of hydrological data. Access to hydrological data pertaining to the classified river basins in India should be eased and made transparent, at least to its own public, for more rigorous and in-depth river research.

183

Chapter 6

Summary and Conclusions

Dam construction has slowed down in the developed countries as most of the technologically and economically viable hydro sites have been tapped (Beck et al., 2012). Countries like the United States have even progressed towards decommissioning of dams to avoid expensive upgrades and address the rising awareness of their detrimental effects (Oliver, 2017). Removal of dams is being increasingly considered for projects that have exceeded their lifetimes and no longer fulfill the purpose for which they were built, but merely pose safety hazards (Katopodis and Aadland, 2014). However, dam construction mainly for hydropower generation is still advancing at a rapid pace in the ‘developing countries and emerging economies’ of Southeast Asia, South America and Africa (Zarfl et al., 2015). Zarfl et al. (2015) stated that the global total of upcoming hydropower projects as of March 2014 amounted to 3700 dams46, out of which 17% were under construction and the rest 83% under planning. They state that a vast majority of these (> 40%) lie in countries with low and low-middle income, including India, Pakistan and Democratic Republic of Congo, but excluding China and Brazil. In Asia, the Ganges- Brahmaputra and the Yangtze basins would witness the ‘highest dam construction’ in the near future (Zarfl et al., 2015). Ultimately, there would be very few free-flowing rivers left in this world.

Given the rapid hydropower development in the Brahmaputra Basin, with serious downstream implications, this thesis presented the downstream hydrological, morphological, and socio-economic consequences of the sole operating hydel scheme in Arunachal Pradesh. It is crucial since the mentioned Himalayan state has the highest number of upcoming and planned projects, but has only one operating so far. Assam as a downstream riparian stands to be affected the most from upstream dams in not just Arunachal Pradesh, but also upon the Brahmaputra (Yarlung Tsangpo) in the Tibetan belt. Both states are also caught in the ‘water grab race’ between India and China (Vidal, 2013), as well as the hydro-politics with Bhutan and Bangladesh.

46 Greater than 1 MW installed capacity (Zarfl et al., 2015).

185

Chapter 6

6.1 Summary of thesis findings

Chapter 2 analyzed and discussed the alterations in flow of the impounded and flow- deprived river - Ranganadi, downstream from the diversion dam. Comparing flows between the unaltered pre-dam period and the altered post-dam period, the changes were evaluated using the Indicators of Hydrological Alteration model of The Nature Conservancy. Post-dam median annual flow in Ranganadi has reduced by 63%. The annual medians post-2002 vary within a much lower range of 7 m3/s (2013) and 94 m3/s (2005), as compared to the annual medians of the pre-dam years. All 22 parameters of the IHA Groups 1 and 2 (i.e., the median monthly flows and annual extreme water conditions respectively) exhibited significant decrease. The difference between the magnitude of pre-dam and post-dam medians in Group 1 ranged between - 46% (in August) and - 80% (in April), while that in Group 2 ranged between - 30% (1-day maximum) and - 88% (1- and 3-day minimum).

There has not been much temporal change in overall channel sinuosity, but braiding exhibited a gradual decrease, especially during the post-dam years of 2004, 2011 and 2014. Certain reaches exhibited more changes in the channel pattern than the river as a whole. Reach 5 displayed an increase in sinuosity from 2002 onwards.

The river exhibited a narrowing pattern with respect to both bankfull and wetted channel widths. Post-dam bankfull channel had reduced by 24%, while the wetted channel reduced by 38%. Width changes were not uniformly distributed along the examined length of the river, and were influenced by formation and abandonment of side channels, avulsions, and unstable bar formations. Reaches 3, 4 and 5 displayed notable post-dam decrease particularly post-2004.

Significant positive correlations were found between channel width (average bankfull, average wetted and maximum wetted widths) and flow (annual mean, pre- monsoon, and monsoonal flows). Although the wetted width (average and maximum both) correlated significantly with the annual maximum flow, the bankfull channel width did not. The maximum bankfull width, on the other hand, did not relate significantly to any of the flow types.

The average rate of bankline migration between the pre-dam (1987-2002) and post-dam (2002-2014) periods did not vary significantly. Migration, erosion and

186

Chapter 6

deposition rates and area differed between the two banks (during both periods). Banklines along specific segments of the river underwent larger shifts, while at other places remained stable. This applied to both the overall pre-dam and post-dam time-scales as well as the shorter periods. Besides the shift made by the avulsion polygon, the river stretch that was subject to frequent bankline shifting was from section R5 to R20 encompassing reaches 2, 3 and 4. The periodic rates of bankline migration (along both banks) were notably larger during 1991-1995 and 2002-2006 due to the 1991 avulsion and course change in the primary channel below R26. Other than that, the migration rates fluctuated across the six time-periods devoid of a distinct pattern of increase or decrease. It can be suggested that the flow alterations made by the hydel project have not significantly affected the patterns of bankline migration, erosion and deposition in the downstream Ranganadi floodplains.

Downstream channel adjustments in the form of narrowing and abandonment due to severe reduction in post-dam flows have affected the conveyance capacity of the Ranganadi River. The result of this is reflected in the exacerbation of downstream floods and occurrence of flash floods, even under diminished peak flows.

Chapter 3 demonstrated the changes in hydrology of the flow-recipient river - Dikrong, downstream from the RHEP powerhouse, following the addition of diverted water from Ranganadi. Annual median flows increased by 138% in the post-dam period, while the elevation in monthly medians ranged from 32% (in July) and 319% (in April). The magnitude of annual minimum flows increased by more than 100%, 7-day, 30-day and 90-day annual maximum flows increased by more than 15%, while the 1-day and 3- day annual maximum decreased by 10% and 5% respectively. No significant change was detected in the median peak flow between the pre-dam and post-dam periods.

After an initial decrease in overall channel sinuosity from 1973 to 2002, SI stabilized at ~ 1.66. Braiding first increased sharply from 1973 (B = 1.12) to 1991 (B = 1.86) and again displayed a steady rise from 2002 (1.68) to 2010 (1.86). Similar abrupt changes in sinuosity and braiding from 1973 to 1987 were obtained in the reach-specific SI and B values. Dikrong displayed a temporal transformation from a predominant single- channel meandering pattern in 1973 to a more braided and multi-channeled pattern by 2014. Channel braiding also exhibited a downstream progression, spreading to the lower most meandering reaches.

187

Chapter 6

The flow-recipient river demonstrated a widening pattern in the bankfull and wetted channel widths. The post-dam average width was higher by 10% for the bankfull channel and 20% for the wetted channel. However, the increase in channel width was detected from 1990 onwards (pre-dam), suggesting that flow augmentation by RHEP may not be the sole cause of channel widening. Particular sections along the river, such as section D17, widened more than a uniformly spread channel widening pattern. With respect to reach-wise change, the lower reaches of the river (reaches 4 and 5) did display distinct width increase in the post-dam period. Increase in channel width simultaneously resulted in erosion of the adjacent riparian lands.

Overall migration of both banklines was comparatively higher during the post-

dam period (average right bankline Rmb = 6 m/y, left bankline = 7 m/y). The maximum

migration rates (max Rmb) during the post-dam period were also higher for both banklines

(right bank max Rmb = 29 m/y, left bank max Rmb = 36 m/y). Majority of the post-dam changes in position of the banklines were initiated during the pre-dam period and continued after 2002. Predominant channel processes that led to migration were lateral movement of banklines away from each other with erosion of the adjacent lands and river widening (D16 to D17), lateral movement of the whole active channel (D19 to D20) and avulsions and meander neck cut-offs (D33 and D35). The daily fluctuations in river water level during the post-dam period might have aggravated bank erosion. The reach of the river from D15 to D30 was subject to repeated and intense bankline shifting increasing the vulnerability and risks of the populace living in adjacent lands.

There were no significant correlations between the average bankfull width and the various flow categories. Moderate relation was found between the maximum bankfull width and the annual mean (rs = 0.53, p = 0.04) and pre-monsoon mean flow (rs = 0.53, p =0.04). Average wetted width did have significant positive correlation with the annual

mean (rs = 0.6, p = 0.02), pre-monsoon (rs = 0.61, p = 0.02) and post-monsoon flow (rs = 0.6, p = 0.015). Not all annual maximum flows (or bankfull flows) resulted in width changes, except for those that formed flood peaks. The 2008 flood peak (957 m3/s) was accompanied by a direct increase in average bankfull width (at 573 m). Width modifications were localized and section-specific as some sections underwent greater changes than others.

188

Chapter 6

While downstream flows in Dikrong have changed significantly following flow- addition from Ranganadi, the channel adjustments over the same time-scale (such as bankline migration) display fluctuations and lack a clear post-dam trend. Notable planform changes were initiated prior to flow augmentation that did not change abruptly in the post-dam period. Channel width was the only planform feature that demonstrated a comparatively steady change in the post-dam period, although still with pre-dam beginnings. However, the cascade development of dams upcoming in the Dikrong basin would, in all likelihood, increase instability in the downstream hydrological and morphological alterations.

Following the hydrological and morphological impacts of the Ranganadi hydel project, Chapter 4 analyzed the socio-economic effects of impoundment and water diversion in Ranganadi. The flash floods in 2008 had significant impact upon the quantity and quality of agricultural lands in the left bank villages, prompting changes in the quantity of land leased in. Percentage of households leasing land for sharecropping increased from 16% to 54% in RL1 and from 48% to 60% in RL2 - the villages with the highest change. This pattern was less evident in the right bank villages located across the river that were essentially unaffected by the 2008 floods. However, these villages (right bank) were still witness to past floods such as the one in 2002.

The impacts on agricultural lands brought about changes in percentage of households involved primarily in farming. Farming as a primary livelihood decreased across all six villages in the post-2002. Instead of a complete farm-exit, households adopted farming (primarily rice cropping) as a secondary livelihood while switching to wage labour and non-farm livelihoods as primary livelihoods. Another adaptation strategy to impact on livelihoods was diversification, i.e., adopting multiple livelihoods, thus reducing the risk and increasing resilience. RL1 and RL2 exhibited the highest increase in multiple livelihoods post-2002.

Decrease in both riverine and riparian fish availability post-dam observation was asserted by more than 60% of surveyed households in all villages (except RR6 where it was reported by ~ 38%). Extreme post-dam decrease in river flow during most days alternating with large volumes of high velocity flows (under sudden reservoir releases) were the primary reason cited for the decline in riverine fish. There is a dearth of proper documentation and baseline information of riverine and wetland fish population in the

189

Chapter 6

Ranganadi. As a result, a comprehensive analysis of its change in relation to the dam was beyond the scope of this study.

Other forms of riverine dependence of the riparian communities such as use of river water for drinking and domestic purposes have also decreased post-dam across all villages. However, along with the dam-induced flow reductions, natural social change factors also play a role in this decrease.

More than 90% in the left bank villages held a negative opinion of RHEP, primarily stemming from the 2008 flash flood impacts and > 82% claimed to have suffered in one way or another from the upstream dam. In the right bank, ≥ 49% asserted the negative view. A vast majority across all the villages (> 70%) claimed feeling threatened by the presence of the dam over Ranganadi. The various reasons cited for the negative perception and threat were - fear of dam breakage, increased frequency and severity of floods (mostly forming flash floods post-dam), uncertainty over increase in river water level, fear of embankment breach and unexpected reservoir releases. Sudden water release and resulting flash floods was cited the most. Demands for decommissioning of RHEP have risen following the 2017 flash floods. A single catastrophic event can go a long way in changing the perception of people towards a particular anthropogenic intervention. The connection between RHEP and the downstream flash floods in 2008 and 2017 is an example, which strengthened the negative perceptions towards hydropower development in Arunachal Pradesh. The agitation against the Lower Subansiri project in the neighboring district also influenced the formation of negative perception towards RHEP.

Finally, it was observed that factors such as proximity to the river and location of homestead and agricultural lands for a household, the topography of the riparian area and the presence of traditional livelihoods, governed the riverine dependence and the spatial distribution of impacts from hydrological changes in the Ranganadi.

Chapter 5 presented the comparisons and contrasts in the post-dam hydrological changes between the flow-recipient and the flow-deprived rivers. The two rivers displayed opposite patterns of change in downstream hydrology and morphology following operations of the Ranganadi hydel project. Natural stressors such as heavy precipitation in the upper catchment causing heavy inflows into the reservoir, plus high

190

Chapter 6

sediment loads of the river, further complicates downstream river hazards. In case of Ranganadi, the downstream socio-economic impacts, have been exacerbated by other anthropogenic interventions such as embankments and a lack of floodplain zoning. The socio-economic effects of river damming were also observed to have resulted directly from the first-order hydrological anomalies and unsustainable management of downstream river flow in Ranganadi.

The second half of Chapter 5 discussed the broader issues associated with hydropower development in Arunachal Pradesh and the various controversies therein. The lacunae in policy dealing with environmental impact assessments of river valley and hydroelectric projects in India were elaborated. Majority of the projects have been categorized as run-of-river, despite differing significantly in size and installed capacity. The “MoU virus” affecting the Arunachal Pradesh government has witnessed grants to projects that lack thorough environmental impact assessments. The downstream impacts still continue to be largely ignored with no negotiating power for the downstream riparian.

Thus, this thesis is a comprehensive and extensive study on the complex responses of river systems to dam regulations, specifically for those that involve water diversion. The Ranganadi hydel project is a much smaller scheme as compared to other upcoming large and mega hydel projects in the region such as the 2000 MW Lower Subansiri, 1750 MW Lower Demwe and the 2880 MW Dibang Multipurpose projects, all located in Arunachal Pradesh. If a project of RHEP’s size can cause such hydrologic, morphologic and socio-economic impacts in the downstream riparian, then how much more would be the impact of these larger projects? The cumulative impacts of the Himalayan dams on a geologically, ecologically and socio-culturally sensitive region (i.e., the Northeast region) would be quite complex and serious. Harnessing the hydro potential of the Himalayan resources is crucial towards improving the electricity situation in India, yet it is also necessary to critically review the impacts of these large projects, especially their complex downstream impacts that have been greatly neglected and under-studied till date within the Indian context. Thus, this thesis is an important contribution towards filling this gap in literature as well as a baseline reference of the probable downstream dam impacts in an economically and ecologically sensitive region, which is also the major source of hydropower in the country. The need for recognizing that the downstream riparian

191

Chapter 6

stakeholders in hydropower projects also suffer from upstream dam interventions has been stressed and recommended in this thesis.

6.2 The way forward

A hydropower project may or may not change the riparian system over and above the natural processes of change, but unless and until the changes are investigated and analyzed, it could be misleading to draw conclusions based only on popular debate. In this thesis, one can observe that contrary to the popular downstream opinion that floods have increased in the Ranganadi post-dam construction, in reality neither the frequency nor the magnitude of post-dam floods are higher than before. However, the key and important change that has occurred and which played a pivotal role in forming of negative perception was the severity and damage left behind by floods that take the form of flash floods in the post-dam period. The nature of flooding has changed due to unsustainable dam regulations and the last-minute approach of RHEP, which has challenged traditional knowledge systems and coping mechanisms of the downstream populations with regards to annual floods. People can no longer rely on the predictability of pre-dam natural flooding patterns and are caught unaware and unprepared by post-dam flash floods. As a result, the loss and damages are higher than before, when there was no dam.

Large hydro projects will happen and be completed in the Brahmaputra Basin. The solution lies in their proper management, especially if inter-basin water diversion is involved. Despite being RoR, the Theun-Hinboun dam in Laos had serious impacts on riparian livelihoods because of inter-basin diversion that ‘reconfigured both the Theun- Kading and Hai-Hinboun river ecosystems’ (Hirsch, 2001; Lebel et al., 2005; International Rivers, n.d.). Lebel et al. (2005) also stated that the downstream impacts of the project gave rise to conflict between the upstream and downstream riparians (‘politics of position’), as local-basin interests got ignored in front of national and regional level interests. Similarly, with respect to hydropower development in the Brahmaputra Basin, local interests within the basin should not be suppressed to triumph over the national and international interests. Such an approach would only fuel the opposition against large dams and delay their completion, resulting in large cost overruns.

Therefore, cumulative impact assessments and basin level studies that give comparable weights to both upstream and downstream effects of dams should be a pre-

192

Chapter 6

requisite for hydropower development. The EIA limits can be distinguished on a case-by- case basis, suited to the design, size, and flow regulation pattern of each hydropower project. One of the most important socio-economic services that hydropower projects in the Himalayan region need to provide is sustainable flood control to the exposed downstream floodplains. At the same time, flow releases should also support the traditional floodplain cropping systems and other riparian livelihoods, instead of causing further impoverishment of already vulnerable communities.

There needs to be strict implementations of environmental flow releases in all upcoming schemes, as well as the older projects, such as RHEP discussed in this thesis, that are still operational. Proper and timely warnings of dam releases from RHEP and other such hydel projects, to the remote and high-risk riparian areas are of utmost importance, so that people get time to take precautionary measures that would help minimize the loss to life and property. In an age when mobile phones have become a basic commodity in every household, even in remote villages, dissemination of bulk messages notifying the opening of spillway and dam releases could be an effective and easy way to send out warnings. Such warning mechanisms would at least help people living in close proximity to the affected river get themselves and their livestock to safer grounds. This is particularly crucial during the monsoon season when floods are a common occurrence in Ranganadi and the rest of the Brahmaputra basin.

Therefore, environmental flow assessment for the Ranganadi hydel project Stage I is of utmost importance and urgently needed. Additionally, there is a dearth of sediment data and studies on sediment transport analysis for Ranganadi and Dikrong and in general for the rivers of the Brahmaputra Basin. This is an important area of research most required under the present scenario of anthropogenic interventions in the basin.

193

References

Adams, W.M. (1985). The Downstream Impacts of Dam Construction: A Case Study from Nigeria. Transactions of the Institute of British Geographers, 10(3), 292-302.

Agarwal, P.K., Singh, V.P., & Jain, S. (2007). Hydrology and Water Resources of India. Water Science and Technology Library, Netherlands: Springer, Science and Business Media.

Aher, S.P., Bairagi, S.I., Deshmukh, P.P., & Gaikwad, R.D. (2012). River Change Detection and Bank Erosion Identification using Topographical and Remote Sensing Data. International Journal of Applied Information Systems (IJAIS), 2(3), 1-7.

Ahmed, F. (2017, June 4). A boat tragedy brings to light pathetic state of 4 villages in North Lakhimpur. The Assam Tribune. Retrieved from http://www.assamtribune.com/scripts/mdetails.asp?id=jun0417/state050.

Ahmed, T., Chetia, S.K., Chowdhury, R., & Ali, S. (2011). Status Paper on Rice in Assam. Rice Knowledge Management Portal. Retrieved from http://www.rkmp.co.in/sites/default/files/ris/rice-state- wise/Status%20Paper%20on%20Rice%20in%20North%20East%20India.pdf.

Al Faraz, F. A. M., & Scholz, M. (2015). Impact of upstream anthropogenic river regulation on downstream water availability in transboundary river watersheds. International Journal of Water Resources Development, 31(1), 28-49, doi:10.1080/07900627.2014.924395.

Alrajoula, M.T., Al Zayed, I.S., Elagib, N.A., & Hamdi, M.R. (2016). Hydrological, socio-economic and reservoir alterations of Er Roseires Dam in Sudan. Science of the Total Environment, 566-567, 938-948. http://dx.doi.org/10.1016/j.scitotenv.2016.05.029.

Anderson, D., Moggridge, H., Warren, P., & Shucksmith, J. (2015). The impacts of run- of-river’ hydropower on the physical and ecological condition of rivers. Water and Environment Journal, 29, 268-276. doi:10.1111/wej.12101.

Andrews, J. & Jelly, N. (2007). Energy Science: Principles, Technologies and Impacts. New York: Oxford University Press.

195

References

Arunachal Pradesh hit by MoU virus. (2009, July 17). The Arunachal Times Archives. Retrieved from http://www.arunachaltimes.in/archives/jul09%2018.html.

Assam State Disaster Management Authority. (2017, July 10). Flood Situation as on today i.e. 10 July, 2017 at 5.00 PM. Flood Report, Government of Assam. Retrieved from http://www.asdma.gov.in/pdf/flood_report/2017/Daily_Flood_Report_10.07.2017.pdf.

Bansal, S. (2017, August 11). MoEF for dropping seven hydel projects in Arunachal. http://www.indiawaterportal.org/articles/moef-dropping-seven-hydel-projects-arunachal.

Baruah, S. (2012). Whose River Is It Anyway? Political Economy of Hydropower in the Eastern Himalayas. Economic & Political WEEKLY, XLVII (29), 41-52.

Baruah, S., & Biswas, S.P. (2003). Stemming the Flood, Killing Biodiversity. The Ecologist Asia, 11(1), 40-41.

Beck, M.W., Claassen, A.H., & Hundt, P.J. (2012). Environmental and livelihood impacts of dams: common lessons across development gradients that challenge sustainability. International Journal of River Basin Management, 1-20. doi:10.1080/15715124.2012.656133.

Benn, P.C., & Erskine, W.D. (1994). Complex channel response to flow regulation: Cudgegong River below Windamere Dam, Australia. Applied Geography, 14, 153-168.

Bhadra, A., Choudhury, S. & Kar, D. (2011). Flood Hazard Mapping in Dikrong Basin of Arunachal Pradesh (India). International Journal of Environmental, Chemical, Ecological, Geological and Geophysical Engineering, 5(12), 861-866.

Bhandari, P.B. (2013). Rural livelihood change? Household capital, community resources and livelihood transition. J. Rural Stud., 32. 126-136. doi: 10.1016/j/jrurstud.2013.05.001.

Bhaskar, U. (2015, July 20). India’s per capita electricity consumption touches 1010 kWh. Livemint. Retreived from http://www.livemint.com/Industry/jqvJpYRpSNyldcuUlZrqQM/Indias-per-capita- electricity-consumption-touches-1010-kWh.html.

Bhattachaiyya, N.N., & Bora, A.K. (1997). Floods of the Brahmaputra River in India. Water International, 22(4), 222-229. doi: 10.1080/02508069708686709.

196

References

Bhattacharjee, U. (2013). Dam Planning Under the Spotlight, A GUIDE TO DAM SANCTIONING IN INDIA. International Rivers, 1-33.

Bhattacharya, S.C., & Jana, C. (2009). Renewable energy in India: Historical developments and prospects. Energy, 34, 981-991. doi:10.1016/j.energy.2008.10.017.

Bhuyan, G. (2012). FLOODS AND ITS SOCIO-ECONOMIC CONSEQUENCES IN THE LOWER REACHES OF DIKRONG RIVER (Doctoral dissertation, North-Eastern Hill University). http://hdl.handle.net/10603/169805.

Bilotta, G.S., Burnside, N.G., Gray, J.C., & Orr, H.G. (2016). The Effects of Run-of- River Hydroelectric Power schemes on Fish Community Composition in Temperate Streams and Rivers. PLOS ONE, 11(5): e0154271, 1-15. doi:10.1371/journal.pone.0154271.

Bora, A.K., (2004). Fluvial Geomorphology. In V.P. Singh, N. Sharma, & C.S.P. Ojha (Eds.), The Brahmaputra Basin Water Resources (47, pp. 88-112). Springer Science+Business Media Dordrecht, Kluwer Academic Publishers. doi 10.1007/978-94- 017-0540-0.

Borgohain, P.L., Phukan, S., & Ahuja, D.R. (2019). Downstream channel changes and the likely impacts of flow augmentation by a hydropower project in River Dikrong, India. International Journal of River Basin Management, 17(1), 25-25. https://doi.org/10.1080/15715124.2018.1439497.

Braatne, J.H., Rood, S.B., Goater, L.A., & Blair, C.L. (2008). Analyzing the Impacts of Dams on Riparian Ecosystems: A Review of Research Strategies and Their Relevance to the Snake River Through Hells Canyon. Environmental Management, 41, 267-281.

Brandt, S.A. (2000). Classification of geomorphological effects downstream of dams. Catena, 40, 375-401.

Bryant, R.G., & Gilvear, D.J. (1999). Quantifying geomorphic and riparian land cover changes either side of a large flood event using airborne remote sensing: River Tay, Scotland. Geomorphology, 29, 307-321.

197

References

Bureau of Indian Standards. (1989). Indian Standard GLOSSARY OF TERMS RELATING TO RIVER VALLEY PROJECTS, PART 10 HYDRO-ELECTRIC POWER STATION INCLUDING WATER CONDUCTOR SYSTEM (First Revision, pp. 1-7). New Delhi, India: Author. Retrieved from https://ia801002.us.archive.org/31/items/gov.in.is.4410.10.1988/is.4410.10.1988.pdf.

Cech,T.V. (2010). Principles of Water Resources: History, Development, Management, and Policy. NJ, USA: John Wiley and Sons, Inc. Retrieved from https://books.google.co.in/books?id=A2nJCPPixGQC&pg=PA13&source=gbs_toc_r&ca d=4#v=onepage&q&f=false.

Census of India. (2011). District Census Handbook Lakhimpur. Retrieved from http://www.censusindia.gov.in/2011census/dchb/1808_PART_B_DCHB_LAKHIMPUR. pdf.

Central Electricity Authority. (2008). Hydro Development Plan for 12th five year plan (2012-2017) (pp. 1-29). Retrieved from http://www.cea.nic.in/reports/hydro/hydro_develop_12th_plan.pdf.

Central Electricity Authority. (2012). National Electricity Plan (Volume 1) Generation (pp. 1-152). http://www.indiaenvironmentportal.org.in/files/file/generation_12.pdf.

Central Electricity Authority. (2017). ALL INDIA INSTALLED CAPACITY (IN MW) OF POWER STATIONS. Retrieved from http://cea.nic.in/reports/monthly/installedcapacity/2017/installed_capacity-12.pdf.

Central Electricity Authority. (2017). Status of H.E. Potential Development – Basin-wise. Retrieved from http://www.cea.nic.in/reports/monthly/hydro/2017/hydro_potential_basin- 12.pdf.

Central Electricity Authority. (n.d.) PREPARATION OF PRELIMINARY FEASIBILITY REPORTS Under 50,000 MW Hydroelectric Initiative. Retrieved from http://www.cea.nic.in/reports/others/hydro/hpi/feasibility_report_50kmw_he.pdf.

Central Water Commission. (2012). INTEGRATED HYDROLOGICAL DATA BOOK (NON-CLASSIFIED RIVER BASINS).

198

References

Centre for Inter-Disciplinary Studies of Mountain & Hill Environment, CISHME. (2009). Environmental Impact Assessment for 1750 MW Demwe Lower HE Project, Arunachal Pradesh (pp. 1-18). New Delhi, India: University of Delhi.

Cernea, M.M. (1997). Hydropower Dams and Social Impacts: A Sociological Perspective. Social Assessment Series, Social Development Papers: Paper No.16, (pp.1- 27). Environment Department, World Bank. Retrieved from http://documents.worldbank.org/curated/en/446311468761673943/pdf/multi-page.pdf.

Choudhury, N. (2014). Legality and legitimacy of public involvement in infrastructure planning: observations from hydropower projects in India. Journal of Environmental Planning and Management, 57(2), 297-315. doi: 10.1080/09640568.2012.743879.

Church, M. (1995). GEOMORPHIC RESPONSE TO RIVER FLOW REGULATION: CASE STUDIES AND TIME-SCALES. Regulated Rivers: Research & Management, 11, 3-22.

Csiki, S., & Rhoades, B.L. (2010). Hydraulic and geomorphological effects of run-of- river dams. Progress in Physical Geography, 1-26. doi: 10.1177/0309133310369435.

Curran, J.H., & McTeague, M.L. (2011). Geomorphology and Bank Erosion of the Matanuska River, Southcentral Alaska, U.S. Geological Survey Scientific Investigations Report 2011-5214 (pp.1-64). . Retrieved from https://pubs.usgs.gov/sir/2011/5214/pdf/sir20115214.pdf.

Dabral, P.P., Pandey, A., & Debbarma, S. (2007). Soil loss estimation of the Dikrong River Basin using IRS-1B Liss II satellite data and GIS. JOURNAL OF THE INSTITUTION OF ENGINEERS (INDIA): AGRICULTURAL AND ENGINEERING DIVISION, 88, 44-51.

Dandekar, P. (2012, September). Damaged Rivers, Collapsing Fisheries: Impacts of Dams on riverine fisheries in India. South Asia Network on Dams, Rivers and People. Retrieved from http://www.indiawaterportal.org/sites/indiawaterportal.org/files/impacts_of_dams_on_riv erine_fisheries_in_india.pdf.

199

References

Das, A.K. (2003). The Brahmaputra’s Changing River Ecology. The Ecologist Asia, 11(1), 42-43.

Das, P J., & Ahmed, F. (2005). Downstream Impacts of the Ranganadi Dam with special emphasis on the Hydrological Impact of the Flow Diversion Plan. In M. Menon, & K. Kohli (Eds.), Large Dams for Hydropower in Northeast India: A Dossier (pp. 57-63). New Delhi, India: Kalpavriksh and SANDRP.

Das, P., Chutiya, D., & Hazarika, N. (2009). Adjusting to Floods on the Brahmaputra Plains, Assam, India. Kathmandu, Nepal: International Centre for Integrated Mountain Development and Aaranyak.

Das, T., Gupta, A., & Arunachalam, A. (2010). Biomass and Productivity of Riparian Vegetation along Dikrong River System, Arunachal Pradesh, India. Assam University Journal of Science and Technology: Biological and Environmental Sciences, 6(1), 13-20.

Dey, S. (2014). Fluvial Processes: Meandering and Braiding. In S. Dey (Ed.), Fluvial Hydrodynamics - Hydrodynamic and Sediment Transport Phenomena (pp. 529-562). New Delhi, India: Springer-Verlag Berlin Heidelberg. doi:10.1007/978-3-642-19062-9.

Dhar, O.N., & Nandagi, S. (2000). A Study of Floods in the Brahmaputra Basin in India. International Journal of Climatology, 20, 771-781.

Dharmadhikary, S. (2008, February 17). Massive dam plans for Arunachal. India Together. Retrieved from http://indiatogether.org/arunachal-environment.

Duarah, C.K. (2015, November 20). Flash floods are burying the lands on the India- Bhutan border in silt, Understanding Asia’s Water Crisis. thethirdpole.net, pp. 1-12. Retrieved from https://www.thethirdpole.net/en/2015/11/20/flash-floods-are-burying-the- lands-on-the-india-bhutan-border-in-silt/.

Dutt, K.N., & Datta, N.C. (1976). Gazetteer of India, Assam State, Lakhimpur District (Present Lakhimpur and Dibrugarh Districts) (pp.1-631). Gauhati, Assam: District Gazetteers, Government of Assam.

200

References

Eaton, B.C., Millar, R. G., & Davidson, S. (2010). Channel patterns: Braided, anabranching, and single-thread. Geomorphology, 120, 353-364. doi:10.1016/j.geomorph.2010.04.010.

Egré, D., & Milewski, J.C. (2002). The Diversity of Hydro Power projects. Energy Policy, 30(14), 1225-1230.

Ellis, F. (1998). Household strategies and rural livelihood diversification. Journal of Development Studies, 35(1), 1-38.

Erlewein, A. (2013). Disappearing rivers - The limits of environmental assessment for hydropower in India. Environmental Impact Assessment Review, 43, 135–143. http://dx.doi.org/10.1016/j.eiar.2013.07.002.

Fan, H., He, D., & Wang, H. (2015). Environmental consequences of damming the mainstream Lancang-Mekong River: A review. Earth-Science Reviews, 146, 77-91.

Fantin-Cuz, I., Pedrollo, O., Girard, P., Zeilhofer, P., & Hamilton, S.K. (2015). Effects of a diversion hydropower facility on the hydrological regime of the Correntes River, a tributary to the Pantanal floodplain, Brazil. Journal of Hydrology, 531, 810-820.

Floods affect over 30,000 in Assam. (2016, October 17). Down to Earth, pp. 1-2. Retrieved from http://www.downtoearth.org.in/news/floods-affect-over-30-000-in-assam- 56023.

Frey, G.W., & Linke, D.M. (2002). Hydropower as a sustainable and renewable energy resource meeting global energy challenges in a reasonable way. Energy Policy, 30(14), 1261-1265.

Friend, P.F., & Sinha, R. (1993). Braiding and meandering parameters. In J.L. Best & C.S. Bristow (Eds.), Branded Rivers (75, pp.105-111). London: Geological Society Special Publications.

Galay, V.J. (1983). Causes of River Bed Degradation. Water Resources Research, 19(5), 1057-1090.

Garg, P. (2012). Energy Scenario and Vision 2020 in India. Journal of Sustainable Energy and Environment, 3(1), 7-17.

201

References

Gautam, Y., & Anderson, P. (2016). Rural livelihood diversification and household well- being: Insights from Humla, Nepal. Journal of Rural Studies, 44, 239-249. http://dx.doi.org/10.1016/j.jrurstud.2016.02.001.

Gendaszek, A S., Magirl, C S., & Czuba, C R. (2012). Geomorphic response to flow regulation and channel and floodplain alteration in the gravel-bedded Cedar River, Washington, USA. Geomorphology, 179, 258-268. http://dx.doi.org/10.1016/j.geomorph.2012.08.017.

Ghanekar, N. (2017, August 3). MoEF recommends dropping 7 hydel projects from Kameng River Basin. DNA. Retrieved from http://www.dnaindia.com/india/report-moef- recommends-dropping-7-hydel-projects-from-kameng-river-basin-2520474.

Ghoshal, S., James, L.A, Singer, M.B., & Aalto, R. (2010). Channel and Floodplain Change Analysis over a 100-Year Period: Lower Yuba River, California. Remote Sensing, 2, 1797-1825.

Giardino, J.R, & Lee, A.A., (2011). Rates of Channel Migration on the Brazos River: Final Report. University of North Texas Libraries. Retrieved from http://texashistory.unt.edu/ark:/67531/metapth415066/.

Gilvear, D.J. (2003). Patterns of channel adjustment to impoundment of the Upper River Spey, Scotland (1942-2000). River Research and Applications, 19, 1-15.

Glucker, A.N., Driessen, P.P.J., Kolhoff, A., & Runhaar, H.A.C. (2013). Public participation in environmental impact assessment: why, who and how? Environmental Impact Assessment Review, 43, 104–111. http://dx.doi.org/10.1016/j.eiar.2013.06.003.

Gogoi, C., & Goswami, D.C. (2014). A study on channel migration of the Subansiri river in Assam using remote sensing and GIS technology. Current Science, 106(8), 1113-1120.

Goldsmith, E. (2003). Rivers. An extract from 5000 Days to Save the Planet by Edward Goldsmith, Nicholas Hildyard, Peter Bunyard and Patrick McCully. The Ecologist Asia, 11(1), 15-18.

202

References

Goswami, D.C. (1985). Brahmaputra River, Assam, India: Physiography, basin Denudation, and Channel Aggradation. WATER RESOURCES RESEARCH, 21(7), 959- 978.

Goswami, D.C. (2008). Managing the Wealth and Woes of the River Brahmaputra. Ishani, 2(4), pp. 1-11 Retrieved from http://www.indianfolklore.org/journals/index.php/Ish/article/view/449.

Goswami, D.C., & Das, P.J. (2003). The Brahmaputra River, India: The eco-hydrological context of water use in one of world's most unique river systems. The Ecologist Asia, 11(1), 9-14.

Goswami, U., Sarma, J.N., & Patgiri, A.D. (1999). River channel changes of the Subansiri in Assam, India. Geomorphology, 30, 227-244.

Government of Arunachal Pradesh. (2008). State Mega Hydro Electric Power Policy. Arunachal Pradesh, India: Author. Retrieved from http://www.indiaenvironmentportal.org.in/files/State%20Mega%20Hydro%20Electric%2 0Power%20Policy%202008.pdf

Goyari, P. (2005). Flood Damages and Sustainability of Agriculture in Assam. Economic & Political Weekly, 40(6), 2723-2729. Retrieved from http://www.jstor.org/stable/4416819.

Graf, W.L. (2006). Downstream hydrologic and geomorphic effects of large dams on American rivers. Geomorphology, 72, 336-360.

Grill, G., Lehner, B., Lumsdon, A.E., MacDonald, G.K., Zarfl, C., Liermann, C.R. (2015). An index-based framework for assessing patterns and trends in river fragmentation and flow regulation by global dams at multiple scales. Environmental Research Letters, 10, 1-13. doi:10.1088/1748-9326/10/1/015001.

Gurnell, A.M. (1997). Channel change on the River Dee meanders, 1946-1992 form the analysis of air photographs. Regulated Rivers: Research and Management, 13, 13-26.

Hazarika, S. (1995). Strangers of the Mist: Tales of War and Peace from India’s Northeast. New Delhi: Penguin Books India.

203

References

Hecht, J., & Lacombe, G. (2014). The effects of hydropower dams on the hydrology of the Mekong Basin. State of Knowledge Series 5, pp. 1-14. Vientiane, Lao PDR, CGIAR Research Program on Water, Land and Ecosystems.

Hill, D.P. (2017). The Discursive Politics of Water Management in India: Desecuritising Himalayan River Basins. South Asia: Journal of South Asian Studies. doi:10.1080/00856401.2017.1380576.

Hirsch, P. (2001). GLOBALISATION, REGIONALISATION AND LOCAL VOICES: THE ASIAN DEVELOPMENT BANK AND RE-SCALED POLITICS OF ENVIRONMENT IN THE MEKONG REGION. Singapore journal of Tropical Geography, 22(3), 237-251.

Huber, A., & Joshi, D. (2015). Hydropower, Anti-Politics, and the Opening of New Political Spaces in the Eastern Himalayas. World Development, 76, 13–25. http://dx.doi.org/10.1016/j.worlddev.2015.06.006.

ICOLD. (2011). CONSTITUTION STATUTS. Paris: ICOLD/CIGB. Retrieved from http://www.icold- cigb.net/userfiles/files/CIGB/INSTITUTIONAL_FILES/Constitution2011.pdf.

ICOLD. (n.d.). Number of Dams by Country Members. Retrieved from http://www.icold- cigb.net/article/GB/world_register/general_synthesis/number-of-dams-by-country- members.

IIT-G Professor asked to study Ranganadi floods. (2017, September 21). The Sentinel Assam. Retrieved from http://www.sentinelassam.com/story/main-news/0/iit-g-professor- asked-to-study-ranganadi-floods/2017-11-12/1/320831#.WsSTktRuZIU.

International Energy Agency. (2015). India Energy Outlook: World Energy Outlook Special Report. Paris, France: OECD/IEA.

International Hydropower Association. (1995-2016). A brief history of hydropower. Retrieved from https://www.hydropower.org/a-brief-history-of-hydropower.

International Rivers (n.d.). Theun-Hinboun Dam and Expansion Project. Retrieved from https://www.internationalrivers.org/campaigns/theun-hinboun-dam-and-expansion-project

204

References

International Rivers. (2016, February). Run-of-River Hydro: Green Energy or Greenwash? Retrieved from https://www.internationalrivers.org/blogs/352-6.

International Rivers. (2017, December). Swindling the Mekong: RUN-OF-RIVER HYDRO. Retrieved from http://www.internationalrivers.org/node/16836.

IRG Systems South Asia Pvt. Ltd. (2015). Cumulative Impact and Carrying Capacity Study of Subansiri Sub Basin including Downstream Impacts (1, pp.1- 433). Retrieved from http://www.indiaenvironmentportal.org.in/files/file/Cumulative%20Impact%20and%20C arrying%20Capacity%20Study%20of%20Subansiri%20basin%20-%20Vol.1.pdf.

Iyer, R. (2003). Water: Perspectives, Issues, Concerns. New Delhi: Sage Publications.

Kalita, P. (2016, October). Surprise water release from Bhutan dam floods 2 Assam districts. Times of India, pp.1-2. Retrieved from https://timesofindia.indiatimes.com/city/guwahati/Surprise-water-release-from-Bhutan- dam-floods-2-Assam-districts/articleshow/54837420.cms.

Katopodis, C., & Aadland, L.P. (2006). Effective dam removal and river channel restoration approaches. International Journal of River Basin Management, 4(3), 153-168. http://www.tandfonline.com/action/showCitFormats?doi=10.1080/15715124.2006.96352 85.

Kaushik, G., & Bordoloi, S. (2016). Ichthyofauna of Ranganadi River in Lakhimpur, Assam, India. Check List, 12(2), 1-6.

Kellerhals, R., Church, M., & Davies, L.B. (1979). Morphological effects of interbasin river diversions. Can. J. Civ. Engin., 6, 18-31.

Kesharwani, M.K. (2006). Overview of Small Hydropower Development in the Himalayan Region. In. AHEC (Alternate Hydro Energy Centre), Himalayan Small Hydropower Summit. Dehradun, Uttarakhand: IIT Roorkee.

Kumar, D., & Katoch, S.S. (2014). Sustainability indicators for run of the river (RoR) hydropower projects in hydro rich regions of India. Renewable and Sustainable Energy Reviews, 35, 101–108. http://dx.doi.org/10.1016/j.rser.2014.03.048.

205

References

Kumar, D., & Katoch, S.S. (2016). Environmental sustainability of run of the river hydropower projects: A study from western Himalayan region of India. Renewable

Energy, 93, 599–607. http://dx.doi.org/10.1016/j.renene.2016.03.032.

Kumar, S., Galapapatti, R., & Chatterjee, C. (2015). Evaluation of Shifting Characteristics of River Brahmaputra Between Guwahati to Goalpara Using LANDSAT and IRS-P6 Liss-III Data. Management of Water Related Disasters - Urban and Rural: Proceeding of the Assam Water Conference-2015, 6-7 February, 2015 (pp.54-63). Guwahati, Assam: Water Resources Department. Retrieved from http://awc.org.in/downloads/awcProceeding_2015.pdf.

Lebel, L., Garden, P., & Imamura, M. (2005). The politics of scale, position, and place in the governance of water resources in the Mekong region. Ecology and Society, 10(2), 18. Retrieved from http://www.ecologyandsociety.org/vol10/iss2/art18/.

Leopold, L.B., & Wolman, M.G. (1957). River Channel Patterns: Braided, Meandering and Straight. Geological Survey Professional Paper 282-B. Washington, DC: US Department of the Interior.

Ligon, F.K., Dietrich, W.E., & Trush, W.J. (1995). Downstream Ecological Effects of Dams. BioScience, 45(3), 183-192.

Lord, M.L., Germanoski, D., Allmendinger, N.E. (2009). Fluvial Geomorphology: Monitoring stream systems in response to a changing environment. In: R. Young & L. Norby (Eds.), Geological Monitoring: Boulder, Colorado (pp.69-103). U.S.: Geological Society of America.

Lu, X.X., & Siew, R.Y. (2006). Water discharge and sediment flux changes over the past decades in the Lower Mekong River: possible impacts of the Chinese dams. Hydrology and Earth System Sciences Discussions, Copernicus Publications, 10 (2), 181-195.

Lyngdoh, R. (2017). Neepco defends Arunachal project. The Telegraph. Retrieved from https://www.telegraphindia.com/1170719/jsp/northeast/story_162642.jsp.

Magilligan, F.J., & Nislow, K.H. (2005). Changes in hydrologic regime by dams. Geomorphology, 71, 61-78.

206

References

Mahanta, R., & Das, D. (2017). Flood induced vulnerability to poverty: Evidence from Brahmaputra Valley, Assam, India. International Journal of Disaster Risk Reduction, 24, 451-461. http://dx.doi.org/10.1016/j.ijdrr.2017.04.014.

Matthew, A. (2017, July 21). 2017 NE floods too are man-made disasters. National Herald, pp. 1-5. Retrieved from https://www.nationalheraldindia.com/environment/2017- ne-floods-too-are-man-made-disasters-assam-arunachal-pradesh.

Menon, M., Vagholikar, N., Kohli, K., & Fernandes, A. (2003). Large Dams in the Northeast – A Bright Future? The Ecologist Asia, 11(1), 3-8.

Ministry of Power. (1998). Policy on Hydro Power Development. Delhi, India: Author. Retrieved from http://powermin.nic.in/whats_new/pdf/hydro_power_policy_developmemt.pdf.

Ministry of Power. (2005). National Electricity Policy. New Delhi, India: Author. Retrieved from http://powermin.nic.in/whats_new/national_electricity_policy.htm.

Ministry of Power. (2008a). Hydropower Policy. New Delhi, India: Author. Retrieved from http://powermin.nic.in/whats_new/pdf/new_hydro_policy.pdf.

Ministry of Power. (2008b, July 3). Press Note - Recent occurrence of flood on 13th /14th June, 2008 in Lakhimpur district of Assam. Retrieved from http://www.pib.nic.in/newsite/erelcontent.aspx?relid=40042.

Ministry of Power. (2018, September 24). FAQs on Hydropower. Retrieved from https://powermin.nic.in/en/content/faqs-hydropower.

Minsitry of Environment and Forests. (2006, September 14). Environmental Impact Assessment Notification, 2006. Retrieved from http://envfor.nic.in/sites/default/files/so1533_4.pdf.

Mitra, A.K. (2004). Flood Management. In V.P. Singh, N. Sharma, & C.S.P. Ojha (Eds.), The Brahmaputra Basin Water Resources (47, pp. 535-560). Springer Science+Business Media Dordrecht: Kluwer Academic Publishers. doi 10.1007/978-94-017-0540-0.

207

References

Mount, N.J., Tate, N.J., Sarker, M.H., & Thorne, C.R. (2013). Evolutionary, multi-scale analysis of river bank line retreat using continuous wavelet transforms: Jamuna River, Bangladesh. Geomorphology, 183, 82-95.

Nathan, H.S.K., Kulkarni, S.S., & Ahuja, D.R. (2013). Pipeline politics—A study of India′s proposed cross border gas projects. Energy Policy, 62, 145-156. http://dx.doi.org/10.1016/j.enpol.2013.06.073.

NEEPCO dam water cause floods in Lakhimpur. (2017, July 10). The Assam Tribune. Retrieved from http://www.assamtribune.com/scripts/detailsnew.asp?id=jul1017/state050.

NEEPCO Not To Be Blamed For Flood In Lakhimpur: CMD. (2017). The Shillong Times. Retrieved from http://www.theshillongtimes.com/2017/07/18/neepco-not-to-be- blamed-for-flood-in-lakhimpur-cmd/.

Neog, P., Sarma, P.K., Chary, G.R., Dutta, S., Rajbongshi, R., Sarmah, K., & … Hazarika, G.N. (2016). Building climate resilient agriculture through traditional floating rice in flash flood affected areas of the North bank plains zone of Assam. Indian Journal of Traditional Knowledge, 15(4), 632-638.

Northwest Hydraulics Consultants. (2006). River Flooding and Erosion in Northeast India. Background Paper No. 4 Development and Growth in Northeast India: The Natural Resources, Water and Environment Nexus (pp. 1-45). Washington, D.C., U.S.A: Ministry of Development of North Eastern Region (MoDONER), India and World Bank.

NRSC-ISRO. (2015). A small hydroelectric plant (130 KW) established near Darjeeling in 1887 ushered the beginning of hydroelectric power development in India. Retrieved from http://india-wris.nrsc.gov.in/wrpinfo/index.php?title=Hydro_Electric_Projects.

Okuku, E. O., Bouillon, S., Ochiewo, J. O., Munyi, F., Kiteresi, L. I., & Tole, M. (2015). The impacts of hydropower development on rural livelihood sustenance. International Journal of Water Resources Development, 32(2), 267-285. doi: 10.1080/07900627.2015.1056297.

Oliver, M. (2017). Liberated Rivers: Lessons From 40 Years of Dam Removal. Science FINDINGS, 193, 1-6. Retrieved from https://www.fs.fed.us/pnw/sciencef/scifi193.pdf.

208

References

Olukanni, D.O., & Salami, A. W. (2012). Assessment of Impact of Hydropower Dams Reservoir Outflow on the Downstream River Flood Regime – Nigeria’s Experience. In. Hossein Samadi-Boroujeni (Ed.), Hydropower - Practice and Application (pp. 61-82). InTech. Retrieved from http://www.intechopen.com/books/hydropower-practice-and- application/assessment-of-impact-of-hydropower-dams-reservoir-outflow-on-the- downstream-river-flood-regime-niger.

Oud, E. (2002). The Evolving Context for Hydropower Development. Energy Policy, 30(14), 1215–1223.

Palaniappan, A.B. (2004). Channel processes. In V.P. Singh, N. Sharma, & C.S.P. Ojha (Eds.), The Brahmaputra Basin Water Resources (47, pp. 196-213). Springer Science+Business Media Dordrecht, Kluwer Academic Publishers. doi:10.1007/978-94- 017-0540-0.

Pandey, A., Dabral, P.P., Chowdhary, V.M., & Yadav, N.K. (2007). Landslide Hazard Zonation using Remote Sensing and GIS: a case study of Dikrong river basin, Arunachal Pradesh. India. Environ Geol, 54, 1517-1529.

Panigrahi, J.K., & Amirapu, S. (2012). An assessment of EIA system in India. Environmental Impact Assessment Review, 35, 23-36. doi:10.1016/j.eiar.2012.01.005.

Petts, G E., & Gurnell, A M. (2005). Dams and geomorphology: Research progress and future directions. Geomorphology, 71, 27-47. doi:10.1016/j.geomorph.2004.02.015.

Phukan, A., Goswami, R., Borah, D., Nath, A., & Mahanta, C. (2012). River Bank Erosion and Restoration in the Brahmaputra River in India. The Clarion: Multidisciplinary International Journal, 1(1), 1-7. Retrieved from http://www.theclarion.in/index.php/clarion/article/view/17/19.

Planning Commission. (2006). Integrated Energy Policy: Report of the Expert Committee. New Delhi: Government of India.

Poff, N.L., Allan, J.D., Bain. M.B., Karr, J.R., Prestegard, K.L., Richter, B.D., Sparks, R.E. & Stromberg, J.C. (1997). The Natural Flow Regime. BioScience, 47(11), 769-784.

209

References

Press Trust of India. (2017). Controlled discharged of Ranganadi Dam water minimised floods. Business Standard. Retrieved from http://www.business-standard.com/article/pti- stories/controlled-discharged-of-ranganadi-dam-water-minimised-floods- 117071801430_1.html.

Rahaman, M.M., & Varis, O. (2009). Integrated water management of the Brahmaputra basin: Perspectives and hope for regional development. Natural Resources Forum, 33, 60–75.

Rampini, C. (2016). IMPACTS OF HYDROPOWER DEVELOPMENT ALONG THE BRAHMAPUTRA RIVER IN NORTHEAST INDIA ON THE RESILIENCE OF DOWNSTREAM COMMUNITIES TO CLIMATE CHANGE IMPACTS (Doctoral dissertation, UC Santa Cruz). https://escholarship.org/uc/item/7qv162sc.

Rao, V.V.K. (2006). Hydropower in the Northeast: Potential and Harnessing Analysis. Background Paper 6, Study on Development and Growth in Northeast India: The Natural Resources, Water, and Environment Nexus (pp. 1-55). Washington, D.C., U.S.A: Ministry of Development of North Eastern Region (MoDONER), India, and World Bank.

Rapp, C.F., & Abbe, T. B. (2003). A Framework for Delineating Channel Migration Zones. Ecology Final Draft Publication #03-06-027 (pp. 1-58). Washington State Department of Ecology, Washington State Department of Transportation. Retrieved from https://fortress.wa.gov/ecy/publications/documents/0306027.pdf.

Reservoir of dams. (2008, May 15). Down to Earth, pp.1-6. Retrieved from http://www.downtoearth.org.in/coverage/reservoir-of-dams-4538.

Richter, B. D., Postel, S., Revenga, C., Scudder, T., Lehner, B., Churchill, A. & Chow, M. (2010). Lost in Development’s Shadow: The Downstream Human Consequences of Dams. Water Alternatives, 3(2), 14-42.

Richter, B.D., Baumgartner, J.V., & Braun, D.P. (1997). How much water does a river need? Freshwater Biology, 37, 231-249.

Richter, B.D., Baumgartner, J.V., Powell, J., & Braun, D.P. (1996). A Method for Assessing Hydrologic Alteration within Ecosystems. Conservation Biology, 10(4), 1163- 1174.

210

References

Roberts, T.R. (1995). MEKONG MAINSTREAM HYDROPOWER DAMS: RUN-OF- THE-RIVER OR RUIN-OF-THE-RIVER?. NAT. HIST. BULL. SIAM Soc., 43, 9-19. Retrieved from http://www.thaiscience.info/Journals/Article/NHB/10439294.pdf.

Rosgen, D.L. (1994). A classification of natural rivers. Catena, 22, 169-199.

Rural Volunteers Centre. (n.d.). The Chang Concept, Granaries on Raised Platform. Retrieved from http://dhemaji.nic.in/flood/Chang_Concept.htm.

Saikia, A. (2017a, July 20). Assam floods: How an embankment in Lakhimpur failed, leaving villages devastated and people dead. Retrieved from https://scroll.in/article/844297/assam-floods-how-an-embankment-in-lakhimpur-failed- leaving-villages-devastated-and-people-dead.

Saikia, A. (2017b, July 22). Severity of Assam floods heightens old fears about dams in the Brahmaputra basin. Retrieved from https://scroll.in/article/844509/severity-of-assam- floods-heightens-old-fears-about-dams-in-the-brahmaputra-basin.

Saikia, H. (2012). Political economy of big Dam in North East India. Basic Research Journal of Social and Political Sciences, 1(1), 1-11.

SANDRP. (2010). The Subansiri gold rush. Dams, Rivers & People, 7(11-12), 1-32. Retrieved from https://sandrp.files.wordpress.com/2018/03/drp_dec_2009-jan_2010.pdf.

SANDRP. (2016, October 14). Bhutan’s Kurichu Dam releases foods Assam, again. Retrieved from https://sandrp.in/2016/10/14/bhutans-kurichu-dam-releases-floods-assam- again/.

SANDRP. (2017, July 17). DRP News Bulletin, Northeast Continues To Experience Floods As Disaster Majorly Because Of Man-made Reasons. Retrieved from https://sandrp.in/2017/07/17/drp-news-bulletin-17-july-2017-northeast-continues-to- experience-floods-as-disaster-majorly-because-of-man-made-reasons/#more-2493.

Santos, Z. (1991, May). Small hydropower for electricity generation. Electricity and the Environment, Background papers for Senior Expert Symposium, Helsinki, IAEA- TECDOC-624 (pp. 369-386). Austria: International Atomic Energy Agency. Retrieved from https://www-pub.iaea.org/MTCD/Publications/PDF/te_624_web.pdf.

211

References

Sarkar, A., Garg, R.D., & Sharma, N. (2012). RS-GIS Based Assessment of River Dynamics of Brahmaputra River in India. Journal of Water Resource and Protection, 4, 63-72.

Sarma, J.N. (1998). FLOOD OF THE BRAHMPAUTRA: CAUSES AND CONTROL MEASURES. Conference Paper, 29-38.

Sarma, J.N. (2004). An Overview of the Brahmaputra River System. In Singh, V.P., Sharma, N., Ojha, C.S.P. (Eds.), The Brahmaputra Basin Water Resources (47, pp. 72- 87). Springer Science+Business Media Dordrecht: Kluwer Academic Publishers. DOI 10.1007/978-94-017-0540-0.

Sarma, J.N. (2006). Bank Erosion and Bankline Migration of the Brahmaputra River in Assam During the Twentieth Century. JOURNAL GEOLOGICAL SOCIETY OF INDIA, 68, 1023-1036.

Sarma, J.N. (2014). The Brahamaputra River in Assam: The Outsized Braided Himalayan River. In V.S. Kale (Ed.), Landscapes and Landforms of India, World Geomorphological Landscapes (pp. 165-172). Dordrecht: Springer Science+Business Media.

Sarma, J.N., & Phukan, M.K. (2006). Bank Erosion and Bankline Migration of the Brahmaputra River in Assam During the Twentieth Century. Journal Geological Society of India, 68, 1023-1036.

Saxena, P., & Kumar, A. (2010). HYDROPOWER DEVELOPMENT IN INDIA. Paper presented in IGHEM-2010, 21-23 October 2010, AHEC, IIT Roorkee, India (pp. 1-6). Retrieved from https://www.researchgate.net/publication/228428780_HYDROPOWER_DEVELOPMEN T_IN_INDIA.

Scudder, T. (2005). The Future of Large Dams: dealing with social, environmental, institutional and political costs (pp. 1-374). UK and USA: Earthscan.

Sen, S., Ganguly, S., Das, A., Sen, J., & Dey, S. (2016). Renewable energy scenario in India: Opportunities and challenges. Journal of African Earth Sciences, 122, 25-31.

212

References

Seth, B.L. (2014). India’s Run-of-River Hydro: Ill-Defined, Under-Studied – and Growing at a Fast Clip. International Rivers. Retrieved from https://www.internationalrivers.org/resources/india%E2%80%99s-run-of-river-hydro-ill- defined-under-studied-%E2%80%93-and-growing-at-a-fast-clip-8268.

Shafroth, P.B., Stromberg, J.C., & Patten, D.T. (2002). Riparian Vegetation Response to Altered Disturbance and Stress Regimes. Ecological Applications, 12 (1), 107-123.

Sharma, M.M., 2006. Institutional Framework of River basin Management in Northeast India: History, Experience and the Way Forward. Background Paper 8, Study on Development and Growth in Northeast India: The Natural Resources, Water, and Environment Nexus (pp. 1-36). Washington, D.C., U.S.A: Ministry of Development of North Eastern Region (MoDONER), India, and World Bank

Shields Jr., F.D., Simon, A., & Steffen, L.J. (2000). Reservoir effects on downstream river channel migration. Environmental Conservation, 27(1), 54-66.

Shukla, P.R. (2000). Future Energy Trends and Carbon Mitigation Strategies for India. In P. Audinet, P.R. Shukla & F. Grare (Eds.), India’s Energy: Essays on Sustainable Development (pp. 21-50). New Delhi: Centre De Sciences Humaines.

Singh, V.K. (2017, July 25). Ranganadi floods, yet again. The Telegraph India, pp. 1-2. https://www.telegraphindia.com/1170725/jsp/northeast/story_163688.jsp.

SJVN Limited. (2012). Doimukh Hydro-Electric Project, Arunachal Pradesh. Pre- Feasibility Report. Shimla, India. Retreived from http://environmentclearance.nic.in/writereaddata/Online/TOR/0_0_17_Jan_2015_161746 [email protected].

Slootweg, R., Vanclay, F., & Schooten, M. (2001). Function evaluation as a framework for the integration of social and environmental impact assessment. Impact Assessment and Project Appraisal, 19(1), 19-28.

Smakhtin, V., & Anputhas, M. (2006). An assessment of environmental flow requirements of Indian river basins. IWMI Research Report 107 (pp. 1-34). Colombo, Sri Lanka: International Water Management Institute.

213

References

Sridharan, P.V. (2009). Environmental Aspects of Power Development. Energy Infrastructure: Priorities, Constraints and Strategies for India (pp. 204-257). New Delhi: Oxford University Press.

Surian, N. (1999). Channel changes due to river regulation: the case of The Piave River, Italy. Earth Surface Processes and Landforms, 24, 1135-1151.

Tare, V., Gurjar, S.K., Mohanta, H., Vishal Kapoor, V., Modia, A., Mathur, R.P., & Sinha, R. (2017). Eco-geomorphological approach for environmental flows assessment in monsoon-driven highland rivers: A case study of Upper Ganga, India. Journal of Hydrology: Regional Studies, 13, 110–121. http://dx.doi.org/10.1016/j.ejrh.2017.07.005.

Thakkar, H. (2003). Dams and Floods. Ecologist Asia, 11(1), 10.

Thakur, S. (2017, February 25). 3 State districts brought under double cropping. The Assam Tribune. Retrieved from

http://www.assamtribune.com/scripts/mdetails.asp?id=Feb2517/at056.

The Nature Conservancy. (2009). Indicators of Hydrological Alteration, Version 7.1, User’s manual.

Totten, M. P., Killeen, T. J., & Farrell, T. A. (2010). Non-Dam Alternatives for Delivering Water Services at Least Cost and Risk. Water Alternatives, 3(2), 207-230.

Tripathi, L., Mishra, A.K., Dubey, A.K., Tripathi, C.B., & Baredar, P. (2016). Renewable energy: An overview on its contribution in current energy scenario of India. Renewable and Sustainable Energy Reviews, 60, 226-233.

U.S. Department of Energy. (n.d.). History of Hydropower. Retrieved from https://energy.gov/eere/water/history-hydropower

USGS. (2016). The USGS Water Science School. Retrieved from https://water.usgs.gov/edu/hyhowworks.html.

Vagholikar, N. (2011). ARE BIG DAMS LEAVING INDIA HIGH AND DRY? Sanctuary Asia, 68-71.

214

References

Vagholikar, N., & Ahmed, M.F. (2003). Tracking a Hydel project. The Ecologist Asia, 11(1), 25-32.

Vagholikar, N., & Das, P. J. (2010). Damming Northeast India. India: Kalpavriksh, Aaranyak and ActionAid. Retrieved from: http://www.kalpavriksh.org/images/EnvironmentandDevelopment/NorthEastDams/Dam ming%20Northeast%20India_Single%20page%20format.pdf.

Verghese, B.G. (2006). Water Resources in the Northeast: Development Options in a Co- operative Framework. Background Paper 1, Study on Development and Growth in Northeast India: The Natural Resources, Water, and Environment Nexus (pp. 1-27). Washington, D.C., U.S.A: Ministry of Development of North Eastern Region (MoDONER), India, and World Bank.

Wade, J.P. (1800). An Account of Assam. R. Sharmah, Madhupur Tea Estate: India.

Wallick, J.R., Grant, G.E., Lancaster, S.T., Bolte, J.P., & Denlinger, R.P. (2007). Patterns and Controls on Historical Channel Change in the Willamette River, Oregon, USA. In A. Gupta (Ed.), Large Rivers: Geomorphology and Management (pp. 491-516). John Wiley & Sons, Ltd.

WAPCOS LIMITED. (2016). CARRYING CAPACITY AND CUMULATIVE IMPACT ASSESSMENT STUDIES FOR HYDROELECTRIC PROJECTS ON THE TRIBUTARIES OF LOHIT RIVER BASIN IN ARUNACHAL PRADESH, VOLUME-I : FINAL REPORT. Ministry of Environment Forest and Climate Change, Government of India.

Williams, G.P., & Wolman, M.G. (1984). Downstream Effects of Dams on Alluvial Rivers. Geological Survey Professional Paper 1286, 1-64. Washington, DC: UNITED STATES GOVERNMENT PRINTING OFFICE.

Winterbottom, S.J. (2000). Medium and short-term channel planform changes on the Rivers Tay and Tummel, Scotland. Geomorphology, 34, 195-208.

Wolman, M.G., & Leopold, L.B. (1957). River Floodplains: Some Observations On Their Formation. US Geological Survey Professional Paper 282-C, 87-109. Washington DC, USA.

215

References

World Bank. (2007). Development and Growth in Northeast India: The Natural Resources, Water and Environment Nexus. Retrieved from http://www.mdoner.gov.in/sites/default/files/silo4_content/REPORT%20STUDIES%20O N%20THE%20NORTH%20EASTERN%20REGION/report%20studies/World_Bank.pdf

World Commission on Dams (WCD). (2000). Dams and Development – A New Framework for Decision- making. London and Sterling: Earthscan Publications Ltd.

Yadama, G.N. (2013). Fires, Fuel, and the Fate of 3 Billion: The State of the Energy Impoverished. USA: OUP.

Zahar, Y., Ghorbel, A. & Albergel, J. (2008). Impacts of large dams on downstream flow conditions of rivers: Aggradation and reduction of the Medjerda channel capacity downstream of the Sidi Salem dam (Tunisia). Journal of Hydrology, 351, 318-330.

Zarfl, C., Lumsdon, A.E., Berlekamp, J., Tydecks, L., & Tockner, K. (2015). A global boom in hydropower dam construction. Aquat Sci, 77, 161-170. doi:10.1007/s00027-014- 0377-0.

216

Appendix A1

-

(12) Chela

Kurung Kumey Kurung 75 - 1430 - - 895 - -

(11) Chomi

Kurung Kurung Kumey Kurung 80 - 1194 - - 1067 -

II -

(10) Oju

Upper Upper Subansiri Subansiri / Singit 1000 4629.93 9979 2825 - 1650 1630

I

-

(9)

Oju

Upper Upper Subansiri Subansiri / Singit 700 3291.58 9827 2825 - 1950 1930

(8) Niare

Upper Upper Subansiri Subansiri / Singit 800 3356.62 11,181 2825 - 1280 1260

(7)

Nalo

Upper Upper Subansiri Subansiri / Singit 360 1732.99 12150 2810 - 765 745

Projects

(6) Naba

Upper Upper Subansiri Subansiri / Singit 1000 3995.25 14,300 2825 - 1035 1022

(5)

Dengser

Upper Upper Subansiri Subansiri / Singit 552 2666.71 17,625 2810 - 630 610

-

(4)

& II Kurung I Kurung Kumey Kurung 330 1435.43 2302 1745.4 310 745 710

(3) Upper Upper Upper Subansiri Subansiri 2000 6768.5 14665 2230 5016 460 420 Subansiri

(2) HEP) Middle (Kamala (Kamala Subansiri Lower Lower Subansiri Kamla 1728 6739 7213 455 430

(1)

Lower Lower Subansiri Lower Lower Subansiri/ Dhemaji Subansiri 2000 7421.59 34,900 2356 12024 205 181

) 2

(MU)

. basin Subansiri the in planned projects hydropower large the of Details

: 1

Features

/s) 3 Appendix A Catchment areaCatchment (km District River capacity Installed (MW) energy Annual 90% in generation year dependable Average annual rainfall Average rainfall annual (mm) Maximum average site discharge dam at (m Full reservoir level (m) (FRL) Minimum Draw Down (m) (MDDL) Level

217

Appendix A1

(12)

------

(11)

------

(10)

12.9 6.55 0.37 800 & 925 Concrete 1655 90 249 11,000 2 126

(9)

31.8 19.7 0.72 700 & 800 Concrete 1955 110 288 10,500 2 126

(8)

15.94 8 0.48 850 & 675 Concrete 1285 100 269 11,500 2 (upper) 126

(7)

163.37 113 2.84 650 & 800 Concrete 770 125 366 12,300 2 (upper) 120

(6)

31.9 23.9 0.81 700 Concrete 1040 110 245 12,000 2 (upper) 142

, 2015.

(5)

89.14 49.48 2.32 650 & 800 Concrete 635 100 383 14,900 2 140

(4)

1075 501 20.25 600 Concrete gravity 750 140 322.5 5500 5 70

(3) 22.2 electric Plants”, Cumulative Impact and Carrying Capacity Study of Subansiri Sub Basin including including Basin Sub Subansiri of Study Capacity Carrying and Impact Cumulative Plants”, electric

- 1755 1010 520 & 520 & 630 Concrete gravity 472 237 533 11,000 2 112

IRG Systems South Asia Pvt. Ltd. Pvt. Asia South Systems IRG

(2)

1927.6 1304.04 33.5 915 to 1315 Concrete gravity 475 216 628 17,416 7 -

(1)

1365 720 33.5 493 to 693 Concrete gravity 210 116 - 37,500 9 -

/s) 3 at FRL at MDDL Design flood (m Spill - way no. Length (m)

) 2

Adapted from “Table 3.4 (a): Salient Features of Hydro of Features Salient (a): 3.4 “Table from Adapted

Downstream Impacts, Final Report, Volume I, by I, Volume Report, Final Impacts, Downstream Gross Storage (Mcum)Gross Area under submergence FRL at (km length tunnel Diversion (m) Type Dam dam of elevation Top (m) above dam of Height foundation deepest level (m) top at dam of Length (m) Spillway

218

Appendix A2

Negligible Negligible Negligible Negligible Negligible Negligible Negligible Negligible Negligible Negligible Negligible Negligible Negligible Villages submerged Negligible Negligible 14 8 1 2 1

-

ctric initiative.

Submer gence (ha) area 104 98 & 2500 1519.76 1296.7 715.53 81.28 200 37.3 48.25 72.3 125 357 3764 281.96 231.56 50.75 166 70 69.56 132

-

UG UG UG UG UG UG UG UG UG UG UG Power house type UG UG UG UG Surface Surface Surface Surface Surface

Annual Energy (Gwh) 16071.6 10823.8 9901 10608.6 4112.4 3995.25 5077.15 4629.93 3356.62 3291.58 3465.9 2345 2345.55 1046.5 2666.71 2247.32 1695.45 2535.8 1716.4 1451.75

type Dam & Earth rockfill concrete concrete concrete concrete concrete concrete concrete concrete concrete concrete concrete concrete concrete concrete concrete rockfill concrete concrete & Earth rockfill

210 241 123 110 125 90 100 110 125 77 108 113 100 105 85 90 75 55 Dam height (m) 155 & 90 200

Live storage (Mcum) & 18.84 16.28 1120 908.57 909.739 34.75 8 57.9 6.35 7.94 12.1 37 - 1177.3 96.55 39.66 7.86 12.56 17.8 5.91 7

MDDL (m) 1030 & 1025 430 828 1054 - 1022 765 1630 1260 1930 980 - 400 1070 610 1340 1260 1315 1430 1660

Full reservoir level (m) 1050 & 1045 490 920 1160 355 1035 805 1650 1280 1950 1020 214 440 1130 630 1360 1270 1355 1440 1670

torage/ Type (S ROR) ROR Storage Storage Storage ROR ROR ROR ROR ROR ROR ROR ROR Storage Storage ROR ROR ROR ROR ROR ROR

Installed capacity (MW) 4000 3000 3000 2600 1120 1000 1000 1000 800 700 700 600 600 600 552 500 500 500 420 400

Subansiri

District valley Dibang Lohit Lohit Lohit Kameng West Subansiri Upper West Siang Subansiri Upper Subansiri Upper Upper West Siang Kameng West Kameng East Kameng West Subansiri Upper valley Dibang valley Dibang West Siang Dibang valley Dibang

: Key features of the hydroelectric schemes planned in Arunachal Pradesh as per the PFRs prepared under the 50000 MW hydroele MW 50000 the under prepared PFRs the per as Pradesh Arunachal in planned schemes hydroelectric the of features Key : I II

2

- -

II

II I

- - - Bhareli Kalai Naying Oju Etalin Demwe Hutong Naba Oju Name of of Name scheme Niare Tato Mihumdon Bhareli Kameng dam Tenga Attunli Emini Hirong Amulin Dengser

Appendix A Sl. No. 1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20

219

Appendix A2

Negligible Negligible Negligible 1 Negligible Negligible Negligible Negligible 4 Negligible Negligible Negligible Negligible Negligible Negligible Negligible Negligible Negligible Negligible Negligible Negligible Negligible

by the Central Central by the

109.5 93.84 238.8 2025 1006 127 5.2 544 565 54 NA 42 11.2 8.33 2.08 6.4 3.09 139.22 31 21.56 25.5 1.7

UG Surface Surface Surface Surface Surface UG Surface Surface UG Surface UG Surface UG Surface Surface UG Surface Surface UG Surface UG

1648.09 1267.38 1733 1435.4 915.5 505 683.66 627.95 551.48 583.14 659.07 748.44 441 401.91 335.26 335.72 359.13 417.82 227.53 174.83 126.45 93.81

ockfill concrete Concrete Concrete Concrete Concrete Concrete Concrete weir Gate R Concrete Concrete Concrete Concrete Concrete Concrete Concrete Concrete Concrete Concrete Concrete Rockfill Concrete

44.5 17 27 19 60 19 22 25 19 95 125 140 50 48 78 18 63 80 70 75 19.5 120

21.8 12 50.37 574 203 18.52 0.78 - 67.21 8.5 - 7 0.43 0.356 0.115 0.108 24.36 1.97 0.89 1.09 0.07

840 1240 745 710 615 685 1670 - 150 1110 - 675 763.4 1187 860.3 1179 1631.1 450 363.5 595 637 816

, (PFRs) REPORTS FEASIBILITY OF PRELIMINARY PREPARATION

860 1250 765 745 640 695 1690 740 165 1130 - 695 767.5 1192.5 865 1185 1637 473.5 367 600 640 822

ROR ROR ROR Storage Storage ROR ROR ROR Storage ROR ROR ROR ROR ROR ROR ROR ROR ROR ROR ROR ROR ROR

390 375 360 330 300 200 165 160 150 150 150 141 120 110 110 100 100 90 80 60 30 30

, Schemes” of PFR Features Salient . (n.d). (n.d). .

Kameng VII:

Dibang valley Dibang Dibang Subansiri Upper Kumey Kurung Kameng East Kameng East valley Dibang Kameng East Papumpare valley Dibang Siang Upper Siang Upper Kameng West Kameng West Kameng East West Kameng West East Saing Kameng East Kameng East valley Dibang Kameng East

II

-

Emra Emra Agoline II I & Kurung Talong Doimukh Nalo Papu Etabue Kapakleyak Elango Ringong Mirak Badao Chanda Pakke Dibbin Utung Simang Sebu Phanchung Ashupani Tarangwarang

Adapted from “Annex - “Annex from Adapted Electricity Authority India , 21 22 23 24 25 26 27 28 29 30 31 32 33 34 35 36 37 38 39 40 41 42

220

Publications and Conferences

Publications

Borgohain, P.L. (2019). Downstream impacts of the Ranganadi hydel project in Brahmaputra Basin, India: Implications for design of future projects. Environmental Development, 30, 114-128. https://doi.org/10.1016/j.envdev.2019.04.005.

Borgohain, P.L., Phukan, S., Ahuja, D.R. (2019). Downstream channel changes and the likely impacts of flow augmentation by a hydropower project in River Dikrong, India. International Journal of River Basin Management, 17(1), 25-35. https://doi.org/10.1080/15715124.2018.1439497.

Conferences

Presented poster titled, “Downstream Hydrologic, Geomorphic and Socio-Economic Impacts of Hydel Projects in Northeast India – Case Studies of the Ranganadi and Dikrong Rivers”, at the 19th International Riversymposium (New Delhi, India), 12th to 14th September, 2016.

Presented poster titled, “Downstream Environmental and Socio-Economic Impacts of The Ranganadi Hydel Project in Northeast India”, at the Manipal University Research Colloquium, 4th to 6th April 2016, Manipal.

221